The Modal and Quasi-Modal Verbs of Obligation and

The Modal and Quasi-Modal Verbs of Obligation andNecessity in Philippine and Singapore English

Kaisa RajalahtiUniversity of Tampere

School of Modern Languages and Translation StudiesEnglish PhilologyPro Gradu Thesis

October 2006

ii

Tampereen yliopistoEnglantilainen filologiaKieli- ja käännöstieteiden laitos

RAJALAHTI, KAISA: The Modal and Quasi-Modal Verbs of Obligation andNecessity in Philippine and Singapore English

Pro gradu -tutkielma, 87 sivua + liitteet (13 sivua)Syksy 2006--------------------------------------------------------------------------------------------------------

Tässä pro gradu -tutkielmassa tarkastellaan Filippiinien ja Singaporen englanninmodaalisia apuverbejä. Keskeiseksi kysymykseksi nousee amerikan- ja brittienglanninvaikutus kyseisiin varieteetteihin, sillä englannin kieli on päätynyt viralliseksi kieleksinäihin Kaakkois-Aasian valtioihin yksinomaan siirtomaavallan vaikutuksesta –Filippiineille Yhdysvaltojen ja Singaporeen Iso-Britannian toimesta. Tutkimuksenkohteena ovat verbit must, should, need, have to, have got to, need to, be supposed to,ought to ja had better.

Tutkielmassa käytetään viittä eri korpusta: Filippiinien, Singaporen ja Iso-Britannianenglantia tutkitaan kutakin The International Corpus of English -korpusten avulla.Nämä korpukset sisältävät sekä puhuttua että kirjoitettua kieltä kultakin alueelta.Amerikanenglannin korpuksia on kaksi, The Santa Barbara Corpus of SpokenAmerican English, joka koostuu puhutusta kielestä, sekä The Freiburg-Brown Corpusof American English, joka koostuu puolestaan kirjoitetusta amerikanenglannista.

Pääpaino on verbien deonttisen modaalisuuden frekvenssillä, mutta episteeminenmodaalisuus otetaan myös huomioon niiden verbien kohdalla, joilla on mahdollistailmaista episteemisiä lauseita. Episteeminen modaalisuus osoittautuu huomattavastiharvinaisemmaksi kuin deonttinen modaalisuus muiden verbien paitsi must kohdalla.Huomioon otetaan myös verbin aikamuodon vaikutus frekvenssiin, sekä subjektinsanaluokka.

Tutkielmassa osoitetaan, että kyseisistä verbeistä eniten käytetyt ovat kaikissatutkituissa varieteeteissa have to, should ja must. Myös verbeillä need to ja supposedto on suhteellisen korkea frekvenssi. Mielenkiintoisimmaksi nousee verbien runsaskäyttö Filippiinien ja Singaporen englannissa verrattuna amerikan- ja brittienglantiin.Tutkielmassa ehdotetaan, että syynä verbien korkeaan frekvenssiin Kaakkois-Aasiassaon aasialaisia kulttuureja leimaava hierarkia ja auktoriteettien totteleminen, joidenvuoksi velvollisuutta ilmaisevien verbien käyttö on sosiaalisesti hyväksyttävämpää.

Avainsanat: apuverbi, Filippiinit, modaalisuus, korpuslingvistiikka, Singapore

iii

Contents1. Introduction ........................................................................................................... 12. Data and Methodology .......................................................................................... 4

2.1. Mood and modality......................................................................................... 52.2. The properties of modal auxiliary verbs .......................................................... 82.3. Root and epistemic meanings.......................................................................... 92.4. Subjectivity and objectivity............................................................................112.5. The International Corpus of English (ICE) .....................................................12

2.5.1. The ICE corpus of Philippine English (ICE-PHI) ....................................142.5.1.1. Aspects of the history of PhiE ..........................................................142.5.1.2. The language situation in the Philippines..........................................16

2.5.2. The ICE Corpus of Singapore English (ICE-SIN)....................................192.5.2.1. Aspects of the history of SinE ..........................................................202.5.2.2. The language situation in Singapore .................................................21

2.5.3. The ICE Corpus of British English (ICE-GB)..........................................242.6. The Santa Barbara Corpus of Spoken American English (SBC) .....................242.7. The Freiburg-Brown Corpus of American English (Frown)............................24

3. Previous Corpus-Based Studies of Modal Verbs in AmE and BrE ........................274. Corpus Findings ...................................................................................................30

4.1. Have to ..........................................................................................................304.1.1. Frequency of have to ...............................................................................31

4.1.1.1. Root and epistemic have to ...............................................................334.1.1.2. Present tense vs. past tense have to ...................................................344.1.1.3. Subjects of have to ...........................................................................37

4.2. Should............................................................................................................394.2.1. Frequency of should ................................................................................41

4.2.1.1. Root and epistemic should and other uses.........................................424.2.1.2. Should have and unfulfilled activity..................................................434.2.1.3. Should not and shouldn’t ..................................................................444.2.1.4. Subjects of should ............................................................................45

4.3. Must...............................................................................................................464.3.1. Frequency of must ...................................................................................48

4.3.1.1. Root and epistemic must ...................................................................494.3.1.2. Must have and must’ve .....................................................................504.3.1.3. Must not and mustn’t ........................................................................514.3.1.4. Subjects of must ...............................................................................51

4.4. Need to ..........................................................................................................524.4.1. Frequency of need to ...............................................................................53

4.4.1.1. Present tense vs. past tense need to ...................................................554.4.1.2. Subjects of need to ...........................................................................57

4.5. Supposed to....................................................................................................584.5.1. Frequency of supposed to ........................................................................59

4.5.1.1. Present tense vs. past tense supposed to ............................................604.5.1.2. Subjects of supposed to ....................................................................64

4.6. Have got to ....................................................................................................654.6.1 Frequency of have got to ..........................................................................66

4.6.1.1. Root and epistemic have got to .........................................................674.6.1.2. Subjects of have got to .....................................................................68

4.7. Ought to.........................................................................................................684.7.1 Frequency of ought to ..............................................................................70

iv

4.7.1.1. Root and epistemic ought to and subjects of ought to........................714.8. Had better......................................................................................................72

4.8.1. Frequency of had better ..........................................................................734.8.1.1. Subjects of had better .......................................................................74

4.9. Need ..............................................................................................................744.9.1. Frequency of need ...................................................................................75

4.9.1.1. Need not and needn’t ........................................................................764.9.1.2. Subjects of need ...............................................................................77

5. Discussion ............................................................................................................785.1. Distribution of the modals and quasi-modals of obligation and necessity .......785.2. Cultural aspects..............................................................................................82

6. Final Remarks ......................................................................................................85References................................................................................................................88Appendix..................................................................................................................92

1

1. Introduction

The influence and power of the English language is unquestionable. It is extensive,

and it is continuously extending its functions and its role as an international lingua

franca as well as a national, regional and local language. From its first spread into new

areas, such as Wales, Scotland, Ireland, The United States, Australia, New Zealand

and South Africa, English has sought and found new ground in East Africa, India, Sri

Lanka, Jamaica, Fiji, and the South-East Asian countries of the Philippines and

Singapore. Despite its indisputable economic and international value, the choice of

English as the official language has been criticised of destroying the national identity

provided by a native language in many of the above mentioned countries.

This study focuses on the English spoken in the Philippines (PhiE) and

Singapore (SinE). English was introduced to the inhabitants of both of these countries

through the medium of education. This choice was motivated by the urge to integrate

the Philippines and Singapore into the western world and international markets, but it

has later been criticised for deteriorating the culture of the native languages of these

countries.

PhiE and SinE have developed into their own varieties on the basis of American

(AmE) and British English (BrE), respectively. The Philippines became the first and

only colony of the United States in 1898, while Singapore was colonised by Britain in

1819. Both countries are now independent and use English as one of their official

languages. The introduction of English to the areas was not an easy task, however, as

the people of both countries speak several other languages. It can therefore be

assumed that the “new” Englishes have not only retained some features of the original

Englishes, but that they have also developed new patterns of use. These patterns are of

2

interest, as the use of PhiE and SinE has been studied fairly little compared to their

donor languages.

Map 1. The Philippines, Singapore and their neighbouring countries(Relief Web 1996)

This study aims to examine the use of PhiE and SinE through modal and quasi-

modal verbs, or more precisely, those of obligation and necessity. The modal and

quasi-modal verbs of obligation and necessity include such verbs as must, have to,

should, ought to, need, had better etc, and they are widely used in both varieties. The

study of these verbs not only offers us interesting information on how English has

been modified in South-East Asia, but also on the power relationships in these two

cultures: the use of a modal of obligation entails authority over the recipient, and the

choice of a certain verb over another has its consequences. Some verbs of obligation

are stronger and more precise than others, and it is therefore possible to make

tentative and preliminary deductions on what sort of obliging is acceptable in a certain

3

culture on the basis of the use of these verbs. To conduct a thorough investigation, the

following research questions will be answered:

1. How are the auxiliary verbs of obligation and necessity used in PhiE andSinE?2. Do these uses differ from those of AmE and BrE?3. If there are differences in the uses, what are they and how can they beexplained?

The linguistic data are collected from five corpora – one from each English-

speaking area mentioned above and two from the USA: The International Corpus of

English (ICE) corpora are used for the study of PhiE, SinE and BrE and The Santa

Barbara Corpus of Spoken American English (SBC) and The Freiburg-Brown Corpus

of American English (Frown) for the study of AmE. The auxiliary verbs will be

presented with their semantic properties and their frequency across the corpora. In

addition, some aspects of the history and the current situation of PhiE and SinE will

be discussed.

4

2. Data and Methodology

An auxiliary verb is a ‘helping verb’, which is used to form tenses, moods and voices

from other verbs (Warner 1993, 1). Modal auxiliary verbs are used to form different

moods, and they express, for example, permitting (e.g. can, may), directing (e.g. must)

and undertaking (e.g. shall) (Huddleston 1988, 79). This study focuses on the

expression of obligation and necessity in PhiE, SinE, AmE and BrE. These two

dimensions of language can be expressed with such modal auxiliaries as must and

should, or with the help of some lexical verbs used as auxiliaries, i.e. quasi-modals,

such as have to and need to. The following illustrates the modals and quasi-modals of

English chosen for the study:

modals quasi-modals

must have to

should have got to

need need to

be supposed to

ought to

had better

To find out how these modals and quasi-modals are used in the varieties, the

following research questions need to be answered: first and foremost, what is the

frequency of the above verbs in these varieties? Which verb is the most frequent, and

which ones are used only marginally? To conduct a thorough investigation, the overall

frequency of the auxiliary verbs in each corpus is taken into consideration. Both

positive and negative counterparts of the verbs have naturally been included and both

full and contracted forms were taken into consideration. Written and spoken corpus

materials are also separated for analysis so that differences between the written and

spoken media can be detected. Most importantly, a closer analysis was conducted to

5

detect differences between the meanings of the modal auxiliaries, i.e. their modalities.

Corpus subcategories were used for the study of modality and the linguistic

environment of the auxiliary verbs (see 2.5.–2.7.). The linguistic environment refers

primarily to the properties of the subject of the verb and its nominal categorisation. As

regards the properties of the verb itself, tense will be the main focus.

The reason for including the study of AmE and BrE along with PhiE and SinE lies

in the history of the Philippines and Singapore. The Philippines was colonised by the

United States and Singapore by Great Britain, and there should therefore be similarities

in the languages spoken in the old and the new countries. There should also be

differences, as the geographical and cultural distance between the Philippines and the

USA, as well as between Singapore and Great Britain, is considerable. To explain the

possible similarities and differences I shall refer to the history of English usage in the

Philippines and Singapore and some cultural aspects.

The data were retrieved from five electronic corpora using Word Smith Tools 4.0

concordance program: the ICE components of PhiE, SinE and BrE, the SBC and the

Frown corpus. Electronic corpora are without a doubt the quickest, the most user-

friendly and the simplest way to retrieve information regarding specific features of

language, as they are extensive and versatile in their number and types of texts. The

ICE corpora contain both spoken and written language, while SBC consists of spoken

language and the Frown corpus of written language only. The corpora are presented

more closely below (see 2.5.–2.7.).

2.1. Mood and modality

To define the semantic nature of modal verbs, it is important to make a distinction

between mood and modality. According to Huddleston (1988, 79–80) the difference

between these is similar to the difference between tense and time, or gender and sex.

6

The clearest distinction can be made according to the opposition between grammar and

semantics in general: mood is a grammatical category of the verb, such as tense, while

modality is a semantic dimension of the statement, such as time. Therefore, mood

traditionally refers to a category which is expressed in verbal morphology (Palmer

1986, 21). Quirk et al. (1985, 149) add that as it is the task of tense to indicate the time

scale of a predication, it is the task of mood to indicate the factual, nonfactual or

counterfactual status of the predication. Indeed, as Nelson (2001, 93) says, mood

essentially refers to “distinctions in the form of a verb phrase that express the speaker’s

attitude towards what is said”.

In English, three types of mood are distinguished: the indicative mood, which is

the unmarked mood used to express statements in general, the imperative mood, which

is used to express directive speech acts such as commands, and the subjunctive mood,

which is used to express wishes and recommendations (Quirk et al. 1985, 149). There

are three sentence types which illustrate the realisation of these categories, namely

declarative (1–2), interrogative (3) and imperative (4) (Quirk et al. 1985, 803):

(1) Pauline gave Tom a digital watch for his birthday. (indicative mood)

(2) I insist that you do this. (subjunctive mood)

(3) Did Pauline give Tom a digital watch for his birthday? (indicativemood)

(4) Listen to me! (imperative mood)

Huddleston and Pullum (2002, 172) note however, that in the case of English, the

question of the relationship of mood and modality is more problematic than in

languages such as Latin, French and German, for example, as mood is not expressed

within the verbal morphology in English. Mood has simply been lost from the

inflectional system by historical change. Huddleston and Pullum argue that the main

7

mood system of English is therefore analytic rather than inflectional, because mood is

marked by the presence or, in contrast, by the absence of distinct modal verbs.

Modality on the other hand refers to the expression of one’s opinion on, or

attitude towards the situation at hand (Declerk 1992, 351). It is primarily a semantic

category, which in English is realised by modal verbs. Huddleston & Pullum (2002,

173) specify that a declarative clause such as he wrote it himself, is regarded as

unmodalised, as the speaker expresses no qualification or special emphasis on the

factuality of the phrase. In contrast, the clause he must have written it himself is

regarded as modalised because the speaker presents the proposition as something that

is not directly known, but as something which is a result of inference.

Palmer adds (1986, 2) that modality does not only relate semantically to the verb

in question, but to the whole sentence. Lyons (1977, 452) argues that although the

scope of modality reaches every part of the sentence, it is nevertheless possible to

distinguish the speaker’s attitude or opinion from the actual proposition of the verb.

Palmer mentions (1986, 14) that in some respect, this comes close to the distinction

between a locutionary act used to say something, and an illocutionary act used to do

something. The actual proposition of the verb then relates to what we say, and

modality relates to what we do with our words – whether we permit, direct or

undertake etc.

The scope of modality is indeed multifaceted. The following illustrates the

multiple attitudes and opinions conveyed by modality and the range of modal verbs

which are used to express them. (Quirk & Greenbaum 1975, 52–57):

(5) ability: He can speak English, but he can’t write it very well. I neverCould play the banjo.

(6) permission: Can/May I smoke in here?

(7) possibility: The road may be blocked. What you say might be true.

8

(8) willingness: He’ll help you if you ask him.

(9) obligation: You should do as he says. You ought to start at once.

(10) logical necessity: They should be home by now. There must be amistake.

The above list is by no means complete as regards the various dimensions of modality.

The interest of this study, however, lies primarily in the uses conveyed by 9–10:

obligation and logical necessity. All of the nine chosen verbs express obligation, but

not all of them (need to, be supposed to, ought to and had better) express logical

necessity.

Because of the lack of morphological marking of mood in English, the

relationship between mood and modality seems to be somewhat vague. I agree with

Palmer (1979, 5) in that although it is possible to describe each modal verb as one of

the above moods of English, the distinction between moods as regards these verbs is

neither very necessary nor useful.

2.2. The properties of modal auxiliary verbs

Modal auxiliary verbs are an extremely complex linguistic phenomenon. Nuyts (2002,

171) says that this complexity is demonstrated in their morphological, syntactic and

semantic behaviour. As the semantic category of modality is the primary interest of

this study, the morphological and syntactic properties of the modals are of secondary

importance. A short examination of the morphology and syntax of modal verbs is

nevertheless in place.

According to Palmer (1968, 21) all modal auxiliary verbs possess four formal

characteristics, namely negation, inversion, code and emphatic affirmation, i.e. NICE-

qualities. Firstly, negation implies that the auxiliary occurs with the negative particle

9

not (11), while the negative counterpart of full verbs is formed with the dummy

auxiliary do (12):

(11) I mustn’t eat sweets anymore.

(12) I don’t love you.

Secondly, inverting the word order also provides proof on whether a verb is an

auxiliary (Palmer 1968, 23). In these constructions, the auxiliary comes before the

subject (13), while a full verb again requires do to complete the inversion (14):

(13) Ought we to ask them?

(14) Do you like chocolate?

The third characteristic, code, refers to the repetition of the auxiliary: this occurs when

a full verb is picked up by an auxiliary later in the sentence (15). Again, Palmer (1968,

24–25) says that do is used with full verbs (16):

(15) We must go and so must you.

(16) I like it and so do they.

Finally, with emphatic affirmation, Palmer (1987, 21) refers to nuclear stress upon the

auxiliary. This is used for affirmation of a doubtful statement or the denial of a

negative:

(17) They should come soon. (‘I am doubtful about their arrival’)

(18) I can come. (‘you are wrong to think I cannot’)

2.3. Root and epistemic meanings

In this study, the semantics the modals of obligation and necessity are studied from the

point of view of root meaning versus epistemic meaning, which brings us back to the

relationship between the proposition and modality – what we say versus what we do.

According to Collins (2005, 251), the root meaning of a modal auxiliary verb of

obligation relates to the social world of duties and directives. This is often called

10

deontic modality1. In other words, deontic modality conveys what is obligatory (19) or

forbidden (20) (Palmer 1979, 3):

(19) You’d better ask him when he comes in. (Palmer 1979, 70)

(20) We shouldn’t be imposing on you in vacation time. (Coates 1983, 63)

In addition, modals are also used to express the speaker's attitude towards the factuality

or the likelihood of the action expressed (Declerk 1992, 351; Huddleston 1988, 78):

epistemic modality is concerned with the speaker’s assumptions or assessment of

possibilities. In fact, as Huddleston (1988, 78) points out, the term epistemic derives

from the Greek word epist meaning ‘knowledge’. As Papafragou (1998, 1) puts it,

epistemic modality has to do with the “degree of speaker commitment to the truth of

the proposition embedded under the modal”. Nuyts (2002, 21) offers a similar, but a

more specific definition. He defines epistemic modality as “(the linguistic expression

of) an evaluation of the chances that a certain hypothetical state of affairs under

consideration (or some aspect of it) will occur, is occurring or has occurred”.

In other words, the speaker judges the event to be real, unreal, certain, probable or

improbable (Bhat 1999, 64). Huddleston and Pullum (2002, 178) in turn state that the

most prototypical difference between root and epistemic modality is that root modality

concerns the attitude to the actualisation of a situation (21), while epistemic modality

concerns the attitude to the factuality of a situation (22–23):

(21) You must pull your socks up.

(22) He must have been delayed.

(23) The next road on the left should be King Street. (Huddleston & Pullum2002, 186)

1 The term deontic was first introduced into modal logic by Georg Henrik von Wright in 1951 (see vonWright 1999, 27)

11

Papafragou (2002, 192) argues that epistemic interpretations presuppose the ability to

reason, and supposedly also the capacity to assess the accuracy of mental

representations. According to Papafragou this is also why epistemic items are parallel

with verbs such as think, know, forget and remember.

Nuyts (2002, 172) adds that only some modals express epistemic modality, and

epistemic usage is therefore secondary to root meaning, at least quantitatively. On the

basis of the results of this study I must agree with Nuyts on this argument, as epistemic

modals were scarce in PhiE, SinE, AmE and BrE, as will be pointed out in chapter 4.

According to Warner (1993, 14) the term root may be used for nonepistemic modals in

general when they are not subcategorised. Therefore, in this study, root will refer to all

nonepistemic tokens.2

2.4. Subjectivity and objectivity

In addition to the categories of root and epistemic modality, other dimensions are also

distinguished when analysing modalised phrases. One of these is subjectivity and

objectivity. According to Collins (2005, 251), the speaker may impose an obligation on

somebody which makes the speaker the source of the obligation. The necessity is then

subjective and internal:

(24) Yeah so you really have to think about it uhm (ICE-PHI S1A-040 [A]: 206)

Huddleston and Pullum (2002, 173) add that although modality is mainly a matter of

the speaker’s attitude, it may apply by extension to the attitudes of those who are

2 Another quite interesting use of the term mood must be noted in connection with deontic and epistemicmodality. Bhat (1999, 8; 63) makes no distinction between mood and modality, and calls the categoriesof modality deontic and epistemic moods. Although it has been stated that the category of mood isalmost non-existent in verbal morphology in English, I believe using these terms interchangeably israther questionable.

12

referred to in the utterance: in Kim thinks he must have written it himself, it is Kim’s

attitude which is indicated, not the attitude of the speaker.

The obligation may also come from outside (Collins 2005, 251). In this case, the

necessity is said to be objective, as the source of obligation is external to the speaker:

(25) Education Minister Tony Tan put his finger on it when, speaking onTuesday night, he urged Singaporeans to ponder, in the cold, clear light ofthe cataclysmic events that have shaken the world, just what the GeneralElection ought to be about. (ICE-SIN W2E-010: 29)

The question of subjectivity and its realisation will be referred to in the accounts of

each modal verb, but their frequency in the case of PhiE, SinE, AmE and BrE is

beyond the scope of this study, as the number of modalised phrases is quite high. I

shall now turn to the sources of these modalised phrases, namely the five electronic

corpora.

2.5. The International Corpus of English (ICE)

The International Corpus of English (ICE) project began in 1990 with the aim of

collecting material of different varieties of English for the benefit of comparative

studies (Greenbaum 1991b, 86; 91). There are research teams around the world

preparing electronic corpora of their own variety of English, and interestingly, the ICE

project is the first systematic investigation of the national variety for most of the

participating countries. In addition to Great Britain, New Zealand and India, of which

such corpora already exist, the countries involved in the project include East Africa

(which includes Kenya, Tanzania and Zambia), Hong Kong, Jamaica and Sri Lanka, to

mention a few (Greenbaum 1991a, 4; 1991b, 86–87). The population represented in the

corpora are adults of at least 18 years of age who have received their education in

English.

13

Each ICE corpus consists of one million words of spoken and written

English which dates from 1990 or later. The text categories and the subcategories are

identical in each ICE corpus, which makes comparison between different varieties easy

(Greenbaum 1996, 5). The corpora include 500 texts of which 300 texts are spoken and

200 texts written English. The number of different text types is presented below

(Greenbaum 1991a, 5; Nelson 1996, 29). The numbers in the brackets indicate the

number of 2 000-word texts in each category:

Spoken(S) (300)

Dialogues(S1) (180)

Private(S1A) (100)

Conversations (90)Phone calls (10)

Public(S1B) (80)

Class Lessons (20)Broadcast Discussions(20)Broadcast Interviews (10)Parliamentary Debates(10)Cross-examinations (10)Business Transactions(10)

Monologues(S2) (120)

Unscripted(S2A) (70)

Commentaries (20)Unscripted Speeches (30)Demonstrations (10)Legal Presentations (10)

Scripted(S2B) (50)

Broadcast News (20)Broadcast Talks (20)Non-broadcast Talks (10)

Written(W) (200)

Non-printed(W1) (50)

StudentWriting

(W1A) (20)

Student Essays (10)Exam Scripts (10)

Letters(W1B) (30)

Social Letters (15)Business Letters (15)

Printed(W2) (150)

Academic(W2A) (40)

Humanities (10)Social Sciences (10)Natural Sciences (10)Technology (10)

Popular(W2B) (40)

Humanities (10)Social Sciences (10)Natural Sciences (10)Technology (10)

Reportage(W2C) (20)

Press reports (20)

14

Instructional(W2D) (20)

Administrative Writing(10)Skills/hobbies (10)

Persuasive(W2E) (10)

Editorials (10)

Creative(W2F) (20)

Novels (20)

Table 1. ICE text categories

Greenbaum (1990, 82) says that the ICE project is descriptive and it will not be

involved in language planning. However, a year later, he added (Greenbaum 1991b, 91)

that the project “will have implications for the teaching of English and in some

countries will be applied to language planning”.

As regards the ICE corpora, two text types were chosen for the closer analysis of

the modality and the subject types of the modals. These are S1A (private dialogue) and

W2F (creative writing). These exact subcategories were chosen because conversation

and fiction are considered the best to exemplify the possible independent developments

in PhiE and SinE compared to AmE and BrE.

2.5.1. The ICE corpus of Philippine English (ICE-PHI)

ICE-PHI was compiled by Dr. Ma. Lourdes S. Bautista, Ms. Jenifer Loy Lising, and Dr.

Danilo T. Dayag of the Department of English and Applied Linguistics at De La Salle

University in Manila (ICE-PHI manual 2004, 3). The corpus was released in 2004.

There is a small exception in the structure of the Philippine corpus as regards the other

ICE corpora: a few spoken texts were not collected and the total number of spoken

texts in ICE-PHI is therefore 278.

2.5.1.1. Aspects of the history of PhiE

The Philippines is the third largest English-speaking country in the world following the

United States and the United Kingdom (Gonzalez 2004, 10). Its culture has been

15

influenced not only by Asian but also by European and American traditions.

Interestingly, SarDesai (1997, 63) says that the degree of European religious and

cultural influence was greater in the Philippines than anywhere else in Asia. Before

Ferdinand Magellan claimed the country for Spain in 1521, the Filipinos traded with

their Chinese and Japanese neighbours. The Spanish brought about the construction of

European buildings and churches as well as the conversion from Islam and belief in

animistic spirits to Christianity (SarDesai 1997, 65; 71). According to Wells (1982,

647), Spanish then became the language of wider communication. However, there was

a succession of revolts against the European colonisers. In 1898 the United States

declared war on Spain, mainly for economic reasons, but also because “the United

States should not only defend but extend the blessings of its Western civilisation to the

region” (SarDesai 1997, 156). The Spanish fleet was destroyed and the Philippine

people thought they had finally won independence. At the end of the war in the Treaty

of Paris in 1898, however, Spain sold the Philippines to the Americans for 20 million

dollars. The rebellion against the colonisers continued, but with little success. In 1934

The Commonwealth of Philippines was established, and the first president, Manuel

Quezon, was given the power to rule some internal affairs.

The Philippines was controlled by the Americans for over forty years, until Japan

occupied the country in 1942 because of their growing interest in Philippine lumber,

hemp and copra (SarDesai 1997, 162–163). The Americans responded with an oil

embargo and in October 1944, the American general Douglas MacArthur began the

liberation of the Philippine people. By the end of the year the country had been cleared

of the Japanese in bloody battles. The republic of the Philippines was proclaimed on 4

July 1946.

16

2.5.1.2. The language situation in the Philippines

The American period introduced English as an official language in the Philippines,

more or less replacing Spanish. The question of language has been contentious in the

Philippines, as the natives speak a variety of languages. As was mentioned in 2.5.1.1.

above, the Philippines is the third largest English-speaking country in the world: 44

million out of the total of 78 million inhabitants speak English (Gonzalez 2004, 10). It

should be noted, however, that these are, for the most part, second language (L2)

speakers, and that the number of native speakers of English in the Philippines is

marginal. English and the Tagalog-based Filipino are the two official languages of the

country. According to McFarland (2004, 65, 66) and Wells (1982, 647), over a half of

the population speaks Tagalog, while the remainder speak such first languages (L1) as

Ilokano, Panganisan and Kapampangan in the north, Hiligaynon, Bikol and Cebuano in

the central islands, and Marano and Magindanao in the south. In fact, over a hundred

languages are spoken in the Philippines (Ledesma 2005, 63). Martin (2004, 255)

argues that Tagalog received its special status during the colonial era because of the

speakers’ concentration around the capital city Manila and because of their influence

on society as administrators, lawyers and politicians. The profusion of languages

inevitably leads to the preference of one language over another, and possibly even to

the extinction of some languages. McFarland (2004, 73) mentions nevertheless that

this may not be such a regrettable thing, as one of the big problems of the Philippines

is the lack of a unifying language.

The English language was introduced in the Philippines through education. The

Spanish colonisers never established a systematic program for education and it was not

until year 1900 that the Americans established public schools. According to Bernardo

(2004, 17–18) The Letter of Instruction to the Philippine Commission, issued by US

17

president William McKinley in 1900 favoured an English-only policy. The original

order was to teach in the native languages of the people, but because there was no

teaching material to be found in those languages, English was chosen as the language

of education. This, according to Bernardo (2004, 18), was done to enable the American

teachers to be more efficient, to unite the Filipinos who spoke a variety of languages,

and finally, quite in keeping with the imperialistic attitude of the Western world, to

provide Filipinos with access to civilisation. The policy attracted early criticism, for

example from the Vice Governor General of the Philippines Islands, George C. Butte,

who insisted in vain in 1931 that children in primary schools be taught in their native

tongue (Bernardo 2004, 18). Education in Filipino was first encouraged in the 1940s,

and in 1953 UNESCO insisted on using the mother tongue of the students to make the

break between home and school less stressful.

The rise of anti-imperialist and anti-American movements in the 1960s

contributed to the establishment of a bilingual policy even further (Bernardo, 2004, 20).

The Bilingual Education Policy of 1974 ratified the use of English and Filipino in

instruction in elementary and high schools: PhiE still remains the only language for the

instruction of science and mathematics, while Filipino is used for other subjects, as

English language competence is considered the best way to ensure economic growth

and access to international markets. Ledesma (2005, 77) notes that while children may

be taught in Filipino, the social status of English becomes indirectly evident to them

through their social environment.

Nevertheless, PhiE is an L2 rather than an L1. According to Hall (1974, xiii) the

English language was nativised in the Philippines. This entails that it was “taken over

by a group of speakers who have previously used some other language, so that the new

language becomes the native language of the group”. Hall (1974, xiv) adds that

18

nativisation does not necessarily involve creolisation, which occurs when two or more

languages converge to form a new indigenous language. Although PhiE is different to

AmE and even though it has adopted features from native Philippine languages, no

extreme reduction or restructuring of English has occurred.

Wells (1982, 647) says that the Filipino are truly multilingual, as they use a

vernacular at home, a lingua franca such as Filipino in transactions with other ethnic

groups, and PhiE in such official situations as business, industry, academia, and wider

communication. Interestingly, according to Otanes and Sibayan (1969, cited in

Gonzalez 2004, 11), prayers and dreams are in the vernacular language of the home,

but letters are written mostly in English. According to Dayag (2004, 41), English is the

dominant language of print media, while broadcast media, such as TV and radio

stations, show a rapid expansion of the domain of Filipino. Indeed, Gonzalez (1991,

360) adds that when a language is learned through a colonial education system, its

domains are often limited, especially when there is a competing indigenous language.

Gonzalez (2004, 12) adds that PhiE is a monostylistic variety: speeches are formal

in style and tone, and there is a tendency to speak as one writes. This may sometimes

create a rather comical effect, if the speaker, for example, refers to him/herself as yours

truly. Platt et al. (1984, 21) add that the pressure for an indigenous national language

was evident after gaining independence. Gonzalez (2004, 12) notes also that the

dominance of English has had a rather powerful effect on Philippine society:

In a Neo-Whorfian belief in the power of language in our perceptions, butlikewise in our attitudes and behaviour, the continuing dominance of anindigenized variety of a post-imperialistic language is considered astumbling block to the jelling of the people into a nation and to perpetuate amentality of dependence which manifests itself … in the dominance of theWestern thought on the thinking of Filipinos and their failure to appreciatetheir own culture….

19

Standard Filipino English is the language of educated Filipinos. It is the type of

English that is acceptable in educated Filipino circles, and it should therefore be

distinguished from creolised forms of English, and the type of English which simply

mixes Tagalog and English (Wells 1982, 647). Indeed, Platt et al. (1984, 147) say that

educated speakers of PhiE use a mixture of Filipino and English, referred to as mix-mix

in less formal situations, even if they could equally well speak English with each other.

In this way, a certain level of informality is achieved where English would be too

formal.

Bolton and Bautista (2004, 3) state that the use of English in the Philippines has

been a source of concern to various sociolinguists: for example, Tollefson (1991, 141)

says that language planning has a powerful impact on how economic resources are

administered and how political power is divided. Tinio (in Bolton & Bautista 2004, 4)

argues that Philippine people are not the only ones that have endured colonialism, but

other countries have managed to survive it with a better mental health, as “they had the

immune system of their own language”. Tupas (2004, 54; 56) adds that the study of

PhiE today should take into consideration the economic and historical unfreedom of

the people who speak it, so that social reform and change can take place. Bolton and

Bautista (2004, 5) state that the challenge remains to provide an education which

supports the use of national and regional languages while taking into consideration the

sociolinguistic realities of the Philippine society, i.e. the need for the opportunity to

receive an education in English.

2.5.2. The ICE Corpus of Singapore English (ICE-SIN)

The ICE-SIN corpus, released in 2002, also consists of one million words of spoken

and written English. The number of the texts as well as the text types is the same as

above (see Table 1.) The ICE research team in Singapore is based at the National

20

University of Singapore (ICE-SIN manual 2002, 3). The team consists of Professor

Paroo Nihilani, Dr Ni Yibin, Dr Anne Pakir and Dr Vincent Ooi.

2.5.2.1. Aspects of the history of SinE

The original inhabitants on the small island at the end of Malay Peninsula were Malay

fishermen, and by the 14th century, Singapore, at the time known as Temasek (‘sea

town’), had become a natural meeting point for Chinese and Indian vessels and even

Portuguese battleships (Turnbull 1977, 2). In the 14th century, the island was also

renamed Singa Pura, (‘lion city’). In the 18th century, Sir Stamford Raffles established

Singapore as a British trading station to protect the empire from the Dutch, who were

becoming stronger in the region (Turnbull 1977, 10).

The population of the country grew rapidly, and free trade attracted merchants

from as far as North America and the Middle East. The Japanese occupied the island in

1942 during World War II, but after the war, Singapore was established as a British

Crown Colony. According to SarDesai (1997, 200), growing nationalism lead to self-

government, but Dixon (2005, 27) contends that it was never the intention of

Singapore to be an independent city-state. It became a part of Malaysia in 1963, but

controversy over the rights of the Chinese citizens in the new united nation lead to

Malaysia’s expulsion of Singapore in 1965, which made the island an independent

republic. MacDougall and Foon (1976, 296) argue that this lead to the abandonment of

the promotion of Malay as the national language.

Today, Singapore is one of the most economically developed societies in the

world. Its economic growth from a quiet fishing village to a technologically modern

city has been more than rapid (Neher 1999, 39). One of the reasons is the country’s

location halfway between India and China, which enabled the nation to become

integrated with the international capitalist system. Neher (1999, 42) adds that location

21

was not the only reason for the rapid expansion of wealth in Singapore: the

overwhelming commitment of the public to continued stability of society, and the

world-famous work ethic have both contributed to the success of Singapore.

2.5.2.2. The language situation in Singapore

English is one of the four official languages of Singapore. According to Singapore

Department of Statistics (2000, in Dixon 2005, 26), 77% of the 4.2 million inhabitants

are Chinese, 14% Malay and 8% Indian. The other official languages are thus

Mandarin Chinese, Malay and Tamil. SinE was an L2 in Singapore for a long period of

time, but it has expanded its functions and become a native or a near-native language

for many speakers (Platt et al. 1984, 22). In fact, according to Dixon (2005, 28), a

bilingual education policy, similar to that in the Philippines (see 2.5.1.2. above) was

passed in 1965, making it obligatory for Singaporean students to study two of the four

official languages: in practice, English and one of the other three, namely Chinese,

Malay or Tamil.

According to Lee (2000, in Dixon 2005, 27) English was encouraged as the

language for interethnic communication. Indeed, Ziegeler (2000, 110) says that in

some respect, SinE has developed in similar ways to a creole, but it is not to be

regarded as one, as it was primarily introduced through the medium of education. This

was a needed approach, as the speakers of Chinese spoke ten different Chinese

“dialects”, such as Hokkien, Teochew and Cantonese, which were not mutually

intelligible (Chua in Dixon 2005, 26; Kuo 1978, 1067). In fact, Platt et al. (1984, 23)

say that the main Chinese dialect spoken is Hokkien, but the official language is

Mandarin, as the government considers it “a more suitable language than any of the

southern Chinese dialects for the expression of Chinese values and culture”.

22

In addition, a minority of the Malay part of the population spoke also Javanese or

Boyanese. Those of Indian descent also spoke a variety of languages from two

language families, namely the Dravidian family (Tamil) and the Indo-European family

(e.g. Punjabi, Gujerati and Bengali). The following table displays the division of ethnic

groups and their home languages in 1990 and 2000. Interestingly, the predominant

home language is in many cases English, even though there were virtually no English-

speaking homes at independence (Singapore Department of Statistics 2000, cited in

Dixon 2005, 30):

Ethnic group / language Overall (%) Overall (%)

1990 2000

Chinese 100.0 100.0

English 19.3 23.9

Mandarin 30.1 45.1

Chinese dialects 50.3 30.7

Other 0.3 0.4

Malays 100.0 100.0

English 6.1 7.9

Malay 93.7 91.6

Other 0.1 0.5

Indians 100.0 100.0

English 32.3 35.6

Malay 14.5 11.6

Tamil 43.2 42.9

Other 10.0 9.9

Table 2. The ethnic groups and their home languages in Singapore

Ho and Platt (1993, 8) say that the features of the languages spoken by these

various ethnic groups have influenced SinE, but the dominant substratum transference

23

comes from Chinese. This concerns syntax, semantics, and to a great extent, phonology.

Ho and Platt continue to note that the influence of Malay is indirect, and works through

more or less pidginised forms of Malay, while Tamil lexical items are not a feature of

SinE, unless when religious festivals or special dishes of food are referred to. The other

Indian languages mentioned have influenced SinE however, as many of the early

English teachers in Singapore were recruited from India and Ceylon (now Sri Lanka).

The transition into English-medium education was first made in subjects like

mathematics and science to ensure the country’s access to international trade and

Western technology. In fact, Dixon (2005, 25) mentions that Singaporean students are

now top-ranked in mathematics and science in worldwide comparison. SinE has grown

gradually in status, but in principle, any of the four official languages can still be

chosen to be used in instruction. He continues to note that the government allocates the

students’ mother tongue based on their ethnicity disregarding their home language.

Platt et al. (1984, 23) say that in reality, however, Tamil-medium education has

vanished.

The success Singapore has experienced in educational outcomes through L2

instruction challenges the view of the pre-eminence of L1 instruction (Dixon 2005, 26).

Kuo (1978, 1069) adds that mass media has also had an important effect on the rise of

English, as it penetrates the lives of a large part of the population. Lim and Foley (2004,

6) say that English has in practice taken over as the language of communication

between ethnic groups, as it is the language of administration, and it thus promises a

higher status and a better career. They also argue that SinE is first and foremost

emerging as the language of the young.

24

2.5.3. The ICE Corpus of British English (ICE-GB)

ICE-GB continues along the lines set for all ICE corpora in the number of words, text

types and dates (see section 2.5.). The ICE-GB research team is based at the Survey of

English Usage in University College, London with Professor Bas Aarts as director.

Released in 1998, it is the oldest one of the ICE corpora used in this study (ICE-GB

manual).

2.6. The Santa Barbara Corpus of Spoken American English (SBC)

The Santa Barbara Corpus of Spoken American English (SBC) parts I-IV will be used

for the study of spoken AmE. Released in 2000–2005, SBC is a part of ICE

representing the American Component. SBC is based on recordings of natural speech

from all over the United States, which represent a variety of people of different regions,

ages, occupations and social backgrounds (Chafe et al. 1991, 71; SBC manual). The

four parts contain 60 speech files and 266 972 words of conversation. This number of

words is based on stripping out all other material but orthographic words from the texts

files. Part I (text files 1–14) was chosen for closer analysis, and it contains 69 557

words.

The difference between the size of SBC and the 600 000 word ICE corpora is

considerable, but SBC is the only spoken AmE corpus available at the moment. The

corpus was compiled by John W. DuBois, Robert Englebretson and associate editors

Wallace L. Chafe, Charles Meyer, Nii Martey and Sandra A. Thompson for the Santa

Barbara Center for the Study of Discourse, University of California.

2.7. The Freiburg-Brown Corpus of American English (Frown)

The well-known Brown Corpus of American English (BC) was compiled in 1964 and

revised in 1979. By the early 1990s, there was a need for an updated version, the kind

25

that would represent the language of the time, and that would also match the

Lancaster-Oslo/Bergen (LOB) corpus as well as the BC. Christian Mair began work on

the Freiburg-Brown corpus (Frown) in 1992, and the one million word corpus was

finally published in 1999. Unlike the ICE corpora, the Frown only includes written

language. The corpus consists of 15 text categories, which are presented below (Frown

manual 1999):

Text Category Numberof texts

A Press: Reportage 44

B Press: Editorial 27

C Press: Review 17

D Religion 17

E Skills, Trades and Hobbies 36

F Popular Lore 48

G Belles Lettres, Biographies,Essays

75

H Miscellaneous 30

J Science 80

K General Fiction 29

L Mystery and DetectiveFiction

24

M Science Fiction 6

N Adventure and Western 29

P Romance and Love Story 29

R Humor 9

Table 3. Frown text categories

The text catgories chosen for closer analysis include K (general fiction), L (mystery

and detective fiction), M (science fiction), N (adventure and western) and P (romance

and love story). These match the creative writing category of the ICE corpora.

26

In the accounts of each verb, SBC and Frown will be referred to together as US.

The spoken part of this self-constructed US corpus refers to SBC and the written part

to Frown.

27

3. Previous Corpus-Based Studies of Modal Verbs in AmE and BrE

In the domain of modal and quasi-modal verbs of obligation and necessity, previous

research is presented in The Longman Grammar of Spoken and Written English

(2000). The grammar includes a corpus study of must, should, have to, had better,

have got to, need to, be supposed to and ought to in AmE and BrE, i.e. all the verbs

under discussion in this study with the exception of modal need. The corpus used, The

Longman Grammar of Spoken and Written English Corpus, (LSGWE) consists of

approximately 5 million words in each register. The registers used in the study include

conversation, fiction, news and academic prose (2000, 25).

The LSWGE study (2000, 488) shows that in conversation, many modals of

obligation and necessity are more common in BrE that in AmE, especially must, had

better and have got to. Should is used in conversation quite frequently in both

varieties. In contrast, have to seems to be more common in AmE. In fact, have to is

the only quasi-modal that is relatively common in conversation as well as in the

written explanatory registers in AmE. In fiction, must and should are more frequent in

BrE than in AmE. Must and should are also common in BrE academic prose. The

study also shows that had better and have got to are infrequent in written language in

both varieties (2000, 489). Need to, ought to and supposed to on the other hand are

less common in all registers.

Peter Collins from the University of New South Wales in Sydney, Australia,

conducted a similar study in 2005 using ICE-GB, the ICE corpora of Australian (ICE-

AUS) and New Zealand English (ICE-NZ) and a self-constructed AmE corpus (C-US)

which consists of 80 000 words from the Frown corpus completed with 29 text files

from parts I and II of SBC (Collins 2005, 252). The modals studied include must,

have to, should, need to, have got to, ought to, need, had better and may/might as well.

28

Have to is the most frequent auxiliary verb in ICE-AUS, ICE-GB and C-US, while

should dominated the ICE-NZ corpus. These are followed by must, need to and have

got to, which are considerably lower in frequency. Ought to, need and had better are

all relatively rare. Unfortunately, supposed to is not included in the study (2005, 253):

0

200

400

600

800

1000

1200

1400

1600

1800

must

have

tosho

uld

need t

o

have

got to

ough

t to need

had b

etter

may/m

ight as

well

ICE-AUSICE-GBICE-NZC-US

Figure 1. Modals and quasi-modals of necessity in Australian, British, New Zealandand American English

Collins (2005, 266–267) attributes the unpopularity of root must compared to

root should and have to to the overt power and hierarchy which marks must: should

and have to are more democratic in this respect as they are not as threatening. Have

got to had a low frequency in the written register, and Collins mentions

colloquialisation as a factor to the decline, as have got to still has a relatively high

frequency in speech. With regard to preferring need to over need, Collins suggests

that the syntactic flexibility of the former supports its use. This lack of flexibility also

undoubtedly accounts for the low frequency of ought to and had better.

29

Collins (2005, 254; 256; 257; 261) adds that epistemic modality is notably more

infrequent than root modality in the case of should, have to and have got to. It is only

in the case of must that epistemic meaning reaches the same frequency as root

meaning. In fact, in both spoken AmE and BrE (2005, 270), epistemic meaning is

more frequent than root meaning (197.5 / 120.2 in AmE and 250.0 / 193.3 in BrE per

one million words). In the written register, on the other hand, root modality is more

frequent in both varieties.

30

4. Corpus Findings

This chapter introduces the semantics of the modal verbs and their frequencies across

the corpora. The verbs will be presented in their order of frequency, from the most

frequent to the least frequent. The frequency of each item is represented in a bar graph

followed by several other illustrations of the properties of the verbs and their

linguistic environment. Where it is mentioned that the numbers represent conversation

and fiction, only the material in the subcorpora chosen has been analysed (see

chapters 2.5., 2.6. and 2.7. for details). The figures have been normalised per one

million words; see the appendix for both raw and normalised figures.

4.1. Have to

According to van Gelderen (2003, 39), modals and to are in complementary

distribution. Therefore, verbal expressions such as have to, have got to, need to and

supposed to cannot be modal. Have to is then classified as a lexical verb used as a

modal, i.e. a quasi-modal.

Palmer (1979, 92) adds that the meaning of have to is simply that the

‘circumstances compel’. Coates (1983, 53; 57) says that the meaning of the verb is

very similar to must, and that it expresses both root (26–27) and epistemic meaning

(28). The latter, however, occurs rarely:

(26) We 've had some rock rocky times you know and uhm some thingsgoing on here but uh I 'm glad to say that we were able to work things outand uh in the end you know we all have to remember that we 're a team.(ICE-PHI S1A-26 [B]: 66)

(27) But Malaysia has to beat Singapore by at least three goals to secure asilver medal. (ICE-SIN S2A-18 [A]: 19)

(28) No one else has read it so it has to be advisory doesn't it?(ICE-GB S1A-68 [B]: 286)

31

Collins (2005, 256) adds that have to has a strong preference for objective root

meaning (29) over subjective root meaning (30):

(29) … they brood their eggs, they hatch their chicks, they brood theirchicks, they feed their chicks, they molt, and then they're out to sea. So allof that has to happen within that six-month period. That’s a pretty busybird. (SBC III-39 [Kirsten]: 615)

(30) Right after we do this the next you'll have to do is that you haveto remember this is to put all these pieces into a basin and you wash themwith detergent or soap. (ICE-SIN S2A-058 [A]: 92)

Interestingly, Coates (1983, 53; 55) contends that root have to is never subjective in

itself, but that it merely expresses the meaning ‘it is necessary for’. In her words, the

authority comes from no particular source, and the meaning is essentially objective.

Collins (2005, 256) says that because of its preference for objective root meaning,

have to is a more attractive option for those who seek a less authoritative modal

expression than, for example, must, which implies strong obligation (see 4.3. below).

Have to also has the past tense form had to, which adds to the versatility of the verb.

4.1.1. Frequency of have to

0

200

400

600

800

1000

1200

1400

1600

1800

2000

ICE-PHI US ICE-SIN ICE-GB

SpokenWritten

Figure 2. Frequency of have to in ICE-PHI, US, ICE-SIN and ICE-GB

32

In addition to first and second person present tense form have to, the third person

present tense form has to and past tense form had to were investigated. The most

likely reason for the high frequency of have to is its less threatening way to express

obligation along with its versatility with regard to tense: the ability to express past and

non-tensed events is unquestionably an important factor. Therefore, have to surpasses

modals must and should in flexibility and in frequency. However, the fact that have to

is quite notably less frequent in written language than in spoken proves that it is still

regarded as somewhat colloquial, or something that belongs to spoken language rather

than written texts.

The difference between spoken PhiE and AmE is quite small (1709 / 1768),

which would indicate that AmE has influenced PhiE more than any other variety or

factor in the use of have to. In written language, however, Filipinos seem to use the

verb more frequently (808 / 582). It is overall surprising that the frequency of have to

is so low in written AmE, i.e. in the Frown corpus, as the Longman study above (see

chapter 3.) suggested that the verb is becoming more acceptable in the written register

in AmE. Even the numbers in written ICE-GB exceed those in the written US corpus.

Surprisingly, SinE shows remarkable independence compared to BrE: have to is

more frequent in ICE-SIN than in ICE-GB in both spoken (1776 / 1310) and written

language (955 / 910). The high frequency of the use of have to in SinE must then be

attributed to something else than the influence of BrE on SinE, perhaps to the

influence of AmE.

33

4.1.1.1. Root and epistemic have to

0

500

1000

1500

2000

2500

3000

ICE-PHI R

OOT

US ROOT

ICE-S

IN R

OOT

ICE-G

B ROOT

ICE-PHI E

PISTEM

IC

US EPIS

TEMIC

ICE-SIN

EPIST

EMIC

ICE-G

B EPISTEMIC

ConversationFiction

Figure 3. Frequency of root and epistemic have toin ICE-PHI, US, ICE-SIN and ICE-GB conversation and fiction

In private conversation, root have to is the most frequent in PhiE: it clearly exceeds

the number in AmE conversation, i.e. the SBC (2418 / 1826). Root have to is also

more frequent in PhiE fiction (1075 / 917), which mirrors the overall situation

between PhiE and AmE regarding the written registers. The situation in SinE and BrE

conversation is no different to spoken language in total, but root have to is slightly

more frequent in BrE fiction than in overall written BrE (1100 / 910).

As was mentioned above (see 4.1.), epistemic have to occurs only rarely. The

examination of PhiE and SinE does not alter this notion. There were only a few

instances of epistemic have to in all of the varieties, and the epistemic use of the verb

was found to be equally infrequent in conversation and in fiction. Interestingly, all of

the epistemic tokens found in the corpora were instances of the form has to. Here are

a few examples:

34

(31) But it has to be a miracle if health is given back to this festering belly.(ICE-PHI W2F-9: 169)

(32) So I called her on the phone she said Oh I assume he has to do it Ialready wrote to him you know (ICE-SIN S1A-15 [A]: 240)

(33) That brings the body count to a total of six young, affluent, well-educated career people. Something has to be in that cocaine, and we haveto find it. (FROWN L-08: 94–96)

4.1.1.2. Present tense vs. past tense have to

0 %

10 %

20 %

30 %

40 %

50 %

60 %

70 %

80 %

90 %

100 %

ICE-PHI S

poken

ICE-PHI W

ritten

US Spok

en

US Writt

en

ICE-SIN

Spo

ken

ICE-SIN

Writt

en

ICE-G

B Spok

en

ICE-G

B Writt

en

Past tensePresent tense

Figure 4. Present tense have to and has to vs. past tense had to in ICE-PHI, US, ICE-SIN and ICE-GB

One of the reasons for the high frequency of have to may be its potential to express

past events. The results of the study show that the speakers of the four varieties take

advantage of this potential. There were several tokens of past tense had to in all

varieties, more so in the written register than the spoken one in every case. This is

most likely due to the reporting nature of written language. The speakers of PhiE and

SinE, however, have not adopted past had to quite as well as the speakers of AmE and

BrE. Platt et al. (1984, 68) say that it is normal in both PhiE and SinE not to make a

35

distinction between present and past events. For example, it is possible to say last

year, I stay three months in Germany. This may be part of the reason for the low

frequency of past tense forms in the case of have to and other verbs below. Platt (1991,

382) adds that there are also differences in past tense marking in SinE according to the

level of education: those who have completed above 4 years of secondary schooling

are more likely to mark verbs for past tense.

The difference between PhiE and AmE is not greater than approximately 10

percentage units in both spoken and written language (14.5% / 25% in spoken, 24.3%

/ 37.6% in written), but the difference is larger between spoken SinE and BrE: (8.6% /

20%). The situation in written language is the opposite: in written SinE, had to is

quite frequent, with a remarkable difference to spoken SinE, while the percentage in

written BrE is lower (29.6% / 21.2%).

CONVERSATION

0

20

40

60

80

100

120

140

do not

have t

o

don't h

ave to

doesn't

have

to

did no

t hav

e to

didn't

have to

have h

ad to

has ha

d to

will hav

e to

would

have to

won't h

ave to

may hav

e to

might h

ave to

ICE-PHIUSICE-SINICE-GB

Figure 5. Positive and negative verb forms of have to in ICE-PHI, US, ICE-SIN andICE-GB conversation

The different forms of have to according to tense and person were not frequent in

conversation or in fiction. This was rather surprising, as the versatility of have to

36

could have been better represented. Full negative forms are nearly non-existent in

speech, and the most frequent forms seem to be the contracted negative forms don’t,

doesn’t and didn’t have to, and the forms will have to and might have to which refer to

future activity. In addition, present and past perfect forms are marginal as well.

However, it must be noted that the number of tokens of these verb forms was quite

low.

In speech, PhiE users seem to use fewer forms of have to compared to AmE

speakers. This could be due to simplification that may have occurred in PhiE,

especially considering the total number of have to, which was higher in ICE-PHI. The

situation between SinE and BrE is more complicated: ICE-SIN had higher frequency

in the case of don’t have to, doesn’t have to and may have to, while ICE-GB had more

tokens of other forms.

FICTION

0

50

100

150

200

250

don't h

ave t

o

doesn

't have

to

did no

t hav

e to

didn't h

ave t

o

have h

ad to

had had

to

will ha

ve to

would ha

ve to

will ha

ve ha

d to

would ha

ve had

to

won't ha

ve to

wouldn't

have to

may ha

ve to

might h

ave t

o


Figure 6. Positive and negative verb forms of have to in ICE-PHI, US, ICE-SIN andICE-GB fiction

The numbers of the verb forms in fiction are quite different to those in

conversation. There are fewer instances of contracted negative forms in all varieties,

37

which is quite expected. The verb forms in PhiE seem to continue along the lines of

spoken language, i.e. there do not seem to be many forms which are frequent in PhiE.

However, in the case of those forms which are used the most (don’t have to, will have

to and would have to) the frequency in ICE-PHI is almost the same as in US.

The frequencies of the different forms of have to in written SinE are low, even

though there are various different forms. These forms, some of which were not found

in ICE-GB, include the perfect forms have had to and would have had to. SinE fiction

may be said, then, to be quite innovative in its use of have to. There were only three

verb forms used in BrE, these were had had to, will have to and would have to, the

last of which had a surprisingly high number of tokens.

4.1.1.3. Subjects of have to

Pers

onal

pro

noun

%

Nou

n %

Inde

finite

pro

noun

%

Rel

ativ

e pr

onou

n %

Non

e %

Det

erm

iner

%

Exis

tent

ial t

here

%

ICE-PHIUS

ICE-SINICE-GB

0,0

10,0

20,0

30,0

40,0

50,0

60,0

70,0

80,0

90,0

100,0


Figure 7. Subjects of have to ICE-PHI, US, ICE-SIN and ICE-GB conversation andfiction

38

PersonalICE-PHI pronoun % Noun % Other % Total %

Conversation 97.7 (N=431) 1.4 (N=6) 1.0 (N=4) 100 (N=441)Fiction 82.2 (N=37) 13.3 (N=6) 4.4 (N=1) 100 (N=45)

USConversation 93.0 (N=119) 3.9 (N=5) 3.1 (N=4) 100 (N=128)

Fiction 74.1 (N=166) 18.8 (N=42) 7.2 (N=16) 100 (N=224)ICE-SIN


ICE-GBConversation 96.4 (N=271) 1.8 (N=5) 1.9 (N=5) 100 (N=281)

Fiction 80.4 (N=37) 17.4 (N=8) 2.2 (N=1) 100 (N=46)

Table 4. Subjects of have to in ICE-PHI, US, ICE-SIN and ICE-GB conversation andfiction

As the frequency of have to is high, there is also variation in the type of the subject.

Personal pronouns were the most frequent subject type of have to in all varieties and

registers. This is not surprising, as obliging is usually directed at a person. The

individual who obliges someone generally has an informal relationship with the

addressee, which indicates that the use of personal pronouns such as you is acceptable.

It can be noted that the speakers of PhiE favour personal pronouns over other subjects

more than speakers of AmE. It could therefore be assumed that PhiE is more restricted

in its use of subjects. In contrast, SinE speakers use mainly personal pronouns as well,

but their use of other subjects is more versatile than that of BrE speakers.

The other types, namely indefinite and relative pronouns, determiners and

existential there were less common. By determiners I refer to all determiners, but

most instances of determiners in the study were demonstratives that and this. There

were also instances where there was no subject, especially in ICE-SIN. These were

mostly cases of ellipsis of the personal pronoun. These are the main reason for the

higher frequency of other subjects in SinE:

39

(34) My father had an accident last year a motorbike accident. So the leggot hurt had to screw uh the bones. (ICE-SIN S1A-077 [A]: 206–207)

Platt (1984, 117–118) and Zhiming (2001, 277)3 say that this phenomenon is not at all

rare in SinE: there is a tendency to imply the subject pronoun rather than state it, as

the subject is usually clear from the context. In fact, Platt also says that replies like

sorry, don’t have are common in Singapore shops.

Personal pronouns as subject were less frequent in fiction. In fact, nouns were

considerably more common as subject in written registers than spoken registers in all

varieties, especially in AmE. There is also more diversity in the choice of subject in

the written corpora, which complies with the notion that written texts are used for

reporting obligations rather than directing them.

4.2. Should

Should is quite a versatile modal verb. According to Papafragou (1998, 3), the root

meaning of the verb encodes tentative necessity, while Coates (1983, 59) adds that “at

its strongest, should takes on the meaning of moral obligation or duty … [and] at its

weakest, it merely offers advice, if subjective, or describes correct procedure, if

objective”. Davis (1977, 102) also says that should indicates that an action is

advisable or desirable.

Huddleston and Pullum (2002, 186) state that should expresses medium strength

modality in its most frequent use. Coates (1983, 58) and Collins (2005, 258) add that

should can be used in four ways: the two most important ones are the root meaning

(35) and the epistemic meaning (36):

(35) Well we should ask him later just to be sure. If not I don't know whatwill happen. (ICE-PHI S1A-012 [C]: 144–144)

3 In fact, Zhiming (2001, 277; 300–303) argues that the evolution of the system of empty categories inSinE is due not only to the influence of the substrate language Chinese, but also to the changes inparameter settings under the pressure of Universal Grammar.

40

(36) That, she thought, should add some excitement, some unpredictabilityinto their lives. (ICE-SIN W2F-002: 60)

Should is also sometimes used as a quasi-subjunctive, and according to Huddleston

and Pullum (2002, 187), this use of should appears with low-degree modality. In the

following examples, should represents mandative, conditional and emotive

subjunctive respectively. In the first type, the term mandative is based on the element

-mand, which is found in such words as demand and mandatory (Huddleston &

Pullum 2002, 174; 995). The meaning of the mandative should then involves the

strong necessity or desirability of an actualisation:

(37) It is perhaps inevitable that the Scottish Arts Council should be heldresponsible for the fact that art galleries in Edinburgh have been closing atan alarming rate. (ICE-GB W2E-006: 27)

Huddleston and Pullum (2002, 186) add that conditional should expresses “slightly

greater doubt than the non-modal counterpart”, i.e. a construction with if. Compare

the following example with the expression if she “accidentally” dropped her pen:

(38) Should she “accidentally” drop her pen, he would surely notice.(ICE-PHI W2F-018:130)

Finally, emotive should can be found in main-clause interrogatives used as rhetorical

questions, or with predicative lexemes which indicate surprise or evaluation

(Huddleston & Pullum 2002, 188):

(39) I don't see why he should mind. (ICE-SIN S1A-062 [A]: 58)

(40) Looking back at all the signals coming out of Moscow in recentmonths, the only surprise about the coup against Mr Mikhail Gorbachev isthat anyone should have been surprised at all. (ICE-SIN W2E-004: 43)

Collins (2005, 259) also says that “should is also used as a first person variant for

hypothetical would, as in I should imagine so, I should say so”:

(41) A fair old walk you know. That 's a good ten minutes I should think.(ICE-GB S1A-023 [A]: 213–214)

41

4.2.1. Frequency of should

0

200

400

600

800

1000

1200

1400

1600


SpokenWritten

Figure 8. Frequency of should in ICE-PHI, US, ICE-SIN and ICE-GB

Should is the second most frequent modal verb in all four corpora. This is most likely

because should is a polite way to express obligation and necessity, i.e. it offers advice

or merely describes correct procedure. In this way, the speaker does not threaten the

independence of his/her interlocutor, nor does a writer necessarily impose ideas on

readers, but merely presents them as favourable options.

Should is slightly more frequent in spoken than written language in both ICE-

PHI (1248 in spoken / 1195 in written) and ICE-SIN (1482 / 1395), but the

differences are so small that it can be stated that should belongs to both spoken and

written language. PhiE in comparison with AmE uses should drastically more. In fact,

should has a surprisingly low frequency in spoken AmE (588). It can then be argued

that the rest of the English-speaking world may have affected the Philippines more

than the USA in the use of should.

SinE seems to have the highest frequency of should in both spoken and written

registers. In contrast, the number of occurrences in spoken BrE is surprisingly low

42

(945), while the number in written BrE (1295) is almost as high as in SinE. The fact

that PhiE and SinE speakers have adopted should into spoken language, while AmE

and BrE speakers have not done so, is an example of colloquialisation in the new

varieties.

4.2.1.1. Root and epistemic should and other uses

0

200

400

600

800

1000

1200

ICE-PHI R

OOT

US ROOT

ICE-SIN

ROOT

ICE-G

B ROOT

ICE-PHI E

PISTEMIC

US EPISTEMIC

ICE-SIN

EPIS

TEMIC

ICE-G

B EPISTEMIC

ICE-P

HI SUBJU

NCTIVE

US SUBJU

NCTIVE

ICE-SIN

SUBJU

NCTIVE

ICE-G

B SUBJUNCTIV

E

ICE-PHI W

OULD

US WOULD

ICE-SIN

WOULD

ICE-G

B WOULD

ConversationFiction

Figure 9. Frequency of should in ICE-PHI, US, ICE-SIN and ICE-GB conversationand fiction

Epistemic should is, as expected, less frequent than root should, but not at all rare.

There were instances of epistemic should in all varieties, mainly in spoken language,

and most of all in SinE (236), followed by BrE (150). The numbers were slightly

lower in PhiE (77) and AmE (86). The other two uses of should, namely the

subjunctive and would are marginal, and the subjunctive seems to be restricted to

written language. These two marginal uses are nevertheless still alive in BrE, but they

have not been adopted into SinE in the same amount.

43

4.2.1.2. Should have and unfulfilled activity

0

50

100

150

200

250

300

350

400


Fiction

Conversation

Figure 10. Should have and should’ve in ICE-PHI, US, ICE-SIN and ICE-GBconversation and fiction

Should has no past tense form, but it is possible to express past obligation in

backshifted environments of shall with the perfect form should have. In fact, Close

(1962, 121) says that should have been suggests unfulfilled activity. This seems not to

be a very popular approach, but there was a considerable amount of tokens

nevertheless. There is no large discrepancy between PhiE and AmE in conversation

(115 / 86) or in fiction (75 / 116). The perfect form is more frequent in PhiE

conversation than in AmE conversation, while the situation is the opposite in fiction.

SinE, on the other hand, shows independence in the high frequency of should have

compared to BrE. The difference is small in conversation (130 / 100), but SinE fiction

seems to favour should have considerably more than written BrE (225 / 150). One

must note, however, that the difference between the actual number of tokens is only 3

(6 tokens in ICE-SIN / 9 tokens in ICE-GB). Here are a couple examples:

(42) Why don't why didn't you just uhm spit it out like you should haverun for th for the ladies ' room. (ICE-PHI S1A-035 [A]: 277)

44

(43) I should have called the police (I've only read about situationslike these in books and watched it on TV), but I know if I do they will justforce me to go outside and testify. (ICE-SIN W2F-007: 147)

4.2.1.3. Should not and shouldn’t

0 20 40 60 80 100 120

ICE-PHI

US

ICE-SIN

ICE-GB

should not %shouldn't %

Figure 11. Should not and shouldn’t in ICE-PHI, US, ICE-SIN and ICE-GB

ICE-PHI should not % shouldn't % Total %Conversation 28.6 (N=6) 71.4 (N=15) 100 (N=21)

Fiction 50 (N=1) 50 (N=1) 100 (N=2)ICE-SIN

Conversation 20.7 (N=6) 79.3 (N=23) 100 (N=29)Fiction 66.7 (N=4) 33.3 (N=2) 100 (N=6)

Table 5. Should not and shouldn’t in ICE-PHI, US, ICE-SIN and ICE-GBconversation and fiction

The contracted shouldn’t was the most frequent negative form in all varieties, but

interestingly, PhiE and SinE seem to favour the full negative form should not more

than AmE and BrE. In fiction the appearance of the full form is hardly surprising, but

there were no instances of should not in AmE and BrE conversation, while there were

a few instances in PhiE and SinE. This may be a sign of more formal attitude in PhiE

and SinE, which was discussed above in the case of PhiE (see 2.5.1.2.):

45

(44) Do you want me to tell him that perhaps he should not pursue you yet?(ICE-PHI S1A-018 [A]: 82)

(45) But they they cautioned me in that I should not uh look at theadministrative side of it so much. (ICE-SIN S1A-032 [B]: 214)

4.2.1.4. Subjects of shouldPe

rson

al p

rono

un %

Nou

n %

Non

e %

Det

erm

iner

%

Rel

ativ

e pr

onou

n %

Exis

tent

ial t

here

%

Inde

finite

pro

noun

%

ICE-PHIUS

ICE-SINICE-GB

0

10

20

30

40

50

60

70

80

90


Figure 12. Subjects of should in ICE-PHI, US, ICE-SIN and ICE-GB







Fiction 82.9 (N=34) 14.6 (N=6) 2.4 (N=1) 100 (N=41)

Table 6. Subjects of should in ICE-PHI, US, ICE-SIN and ICE-GB conversation andfiction

46

The subjects of should are no different to those already discussed in the case of have

to. Personal pronoun is again the most frequent subject type followed by noun, but

other subjects of should seem to be more frequent in conversation than those of have

to in all varieties with the exception of AmE, but it must also be noted that the number

of instances is very low in AmE conversation. SinE speakers, again, use more nouns

as subjects than BrE speakers, but other subjects are more frequent in BrE than in

SinE. Nevertheless, the trend of omitting the subject in SinE conversation, which was

discussed in the case of have to above (see 4.1.1.3. above), continues with should.

In conclusion, fiction is again more prolific in choosing different subjects than

conversation with the exception of other subjects in BrE: these are surprisingly few in

ICE-GB. In contrast, other subjects are just as frequent in SinE fiction as are nouns.

4.3. Must

If a speaker wishes to use an expression with which they can assert power and

authority over the addressee, must is sure to make that impact. Indeed,

Woisetschlaeger (1985, 110) says that must in comparison with should provides a

safer guess, a more comforting reassurance and harsher criticism. Coates (1983, 33)

also adds that a native speaker’s stereotype of the utterer of must is that he/she is

interested in getting the subject to perform the action, and he/she has authority over

the subject.

According to Coates (1983, 31–33) the root meaning of must expresses

obligation or necessity. She goes on to note that the root meaning of must comes close

to an imperative phrase, i.e. I order you to do something, and it therefore implies

strong obligation. Therefore, subjective must is essentially performative – by uttering

the word, the speaker simultaneously tells the addressee to do something:

47

(46) You must get a lawyer for this you know. (ICE-PHI S1B-068 [B]: 148)

(47) Oh yes well you know you must get over it as best as you can. (ICE-GB:S1A-049 [B]: 17)

Meanwhile, epistemic must conveys the speaker’s confidence in the truth of the

utterance, i.e. by uttering epistemic must, the speaker makes a logical inference

(Christophersen & Sandved 1970, 195; Coates 1983, 41). The utterance is based on

deduction from facts which are known to the speaker and on logic. Davis (1977, 102)

says that these facts or this logic may be based on physical laws or rules. According to

Huddleston and Pullum (2002, 181), subjective epistemic must (48) involves

pragmatic weakening, and it will therefore be interpreted as a mere confident

inference, while objective epistemic must (49–50) involves strict semantic necessity:

(48) I think the time-table must have got mixed up. (ICE-SIN S1A-[B]: 63)

(49) Now if you have a transfer from one place to another definitelythere must be some relationship between the two phases and since youhave a interface that equilibrium relationship most probably I say firstmost probably occur at the interface. (ICE-SIN S1B-007 [A]: 45)

(50) I must be the same age as the man who killed the world outside theWalls, and I can recognize him now when I look upon his face. (ICE-PHI W2F-015: 199)

48

4.3.1. Frequency of must

0

200

400

600

800

1000

1200


Spoken

Writ ten

Figure 13. Frequency of must in ICE-PHI, US, ICE-SIN and ICE-GB

As Collins (2005, 253–254, see also chapter 3.) says, the use of must is on the decline,

mainly because of its authoritative nature. Collins suggests that this is due to

“democratisation”, a phenomenon first presented by Myhill (1995, 202), which entails

changes in language induced by social reasons. Myhill argues that must is strongly

associated with the expression of unequal power, and the strive for equality has lead

people to mark individuality and solidarity in more overt ways. This would explain

the recent decline in the frequency of must.

In all four dialects, must is more frequent in written language. Interestingly, both

PhiE and SinE use must more than their colonisers, especially in spoken language.

The reason for its high frequency in the South-East Asian varieties remains unclear.

However, it may be that it is more acceptable in Asia to oblige people than in our

Western culture, and that power relationships are more overtly expressed. This would

explain the large difference between the new and the old Englishes.

49

Must seems to be quite frequent in SinE in both spoken and written registers

(993 / 1068), which is unique because the verb is clearly preferred by written

language in the other varieties. The extensive use of must in written SinE may be due

to the influence of colonisation, as must is quite frequent in written BrE as well.

4.3.1.1. Root and epistemic must

0

100

200

300

400

500

600

700

800

900

ICE-PHI R

OOT

US ROOT

ICE-SIN

ROOT

ICE-G

B ROOT

ICE-PHI E

PISTEMIC

US EPISTEMIC

ICE-SIN

EPISTEMIC

ICE-G

B EPISTEMIC

Conversat ion

Fiction

Figure 14. Root and epistemic must in ICE-PHI, US, ICE-SIN and ICE-GBconversation and fiction

Must is the only modal verb in this study whose epistemic use reaches the same

numbers as the root use: making inferences is a more popular approach than giving

orders. This was also found by Collins (2005, 266–267, see chapter 3.) In fact, in

conversation, epistemic must is more frequent than root must in all dialects with the

exception of SinE. This solidifies the notion that it may be more acceptable to oblige

people to act in Singapore. PhiE speakers on the other hand seem to be less

enthusiastic about the epistemic use than AmE speakers (110 / 201). As regards

fiction, root meaning is only slightly more frequent than epistemic meaning in all

50

corpora except ICE-GB, which has a considerably high frequency of the epistemic use

(775).

4.3.1.2. Must have and must’ve

0

50

100

150

200

250

300

350

400

450

500


Fiction

Conversat ion

Figure 15. Must have and must’ve in ICE-PHI, US, ICE-SIN and ICE-GBconversation and fiction

As all modal verbs, must has no past tense form. Close (1962, 122) says that the

present perfect form must have tells us what has necessarily taken place, i.e. it

expresses necessity. The non-fulfilment of the obligation is simply impossible, as the

obligation was inevitable:

(51) There must have been a build-up of ill-will somewhere along the line.(ICE-PHI S1B-042 [C]: 80)

All the tokens of must have in the corpora are then essentially epistemic. This is most

likely the reason for the high frequency of must have, as it is more frequent than

should have (see 4.2.1.2. above). Must have is nevertheless more restricted to written

language. The frequency in both conversation and fiction in ICE-PHI is again higher

than in US (244 / 163), while the number in ICE-SIN is this time slightly lower than

51

in ICE-GB (385 / 430). The differences in both cases are made by the numbers in

fiction as the numbers in conversation are nearly equal between PhiE and AmE and

between SinE and BrE.

4.3.1.3. Must not and mustn’t

ICE-PHI must not mustn't US must notConversation 0 0 Conversation 0

Fiction 50 (2) 50 (2) Fiction 21 (5)ICE-SIN

Conversation 5 (1) 5 (1) Conversation 0Fiction 0 25 Fiction 50 (2)

Table 7. Must not and mustn’t in ICE-PHI, US, ICE-SIN and ICE-GB conversationand fiction

The frequency of must not and mustn’t was surprisingly low in all corpora. In fact,

there was only one instance of must not in spoken ICE-SIN, and none in the rest of the

corpora. This is most likely again due to the strict and authoritative nature of must not

and its contracted version. Close (1962, 120) also reminds us that the meaning of I

must not go is that ‘I am forbidden to go’, and if one wants to negate the obligation,

need must be used: I need not go.

4.3.1.4. Subjects of must

ICE-PHIPersonal

pronoun % Noun % Other % Total %Conversation 62.9 (N=22) 14.3 (N=5) 22.9 (N=8) 100 (N=35)

Fiction 60 (N=21) 31.4 (N=11) 8.6 (N=3) 100 (N=35)US

Conversation 62.5 (N=10) 25 (N=4) 12.5 (N=2) 100 (N=16)Fiction 70.3 (N=83) 19.5 (N=23) 10.2 (N=12) 100 (N=118)

ICE-SINConversation 67.6 (N=115) 7.6 (N=13) 24.7 (N=42) 100 (N=170)

Fiction 58.8 (N=20) 32.4 (N=11) 8.8 (N=3) 100 (N=34)

52

ICE-GBPersonal

pronoun % Noun % Other % Total %Conversation 79.5 (N=70) 3.4 (N=3) 17.0 (N=15) 100 (N=88)

Fiction 66.7 (N=30) 24.4 (N=11) 8.9 (N=4) 100 (N=45)

Table 8. Subjects of must in ICE-PHI, US, ICE-SIN and ICE-GB conversation andfiction

The subjects of must present no surprises: most of the subjects are personal pronouns

in all dialects. Fiction is again more versatile in the choice of subject. The number of

verb tokens with no subject in spoken SinE has been high in both of the other verbs

discussed (see 4.1.1.3 and 4.2.1.4 above), but in the case of must, it is very high

indeed (19.4%, N=33):

(52) Next time must bring a tumbler of water. (ICE-SIN S1A-036 [B]: 288)

4.4. Need to

Need to is used to express “an attitude towards an unfulfilled activity” (Close 1962,

128). Quasi-modal need to differs semantically from the verbs mentioned above in

that the compulsion implicated comes from within the speaker (Perkins 1983, 62–63).

For example, if it is stated that I need to lie down, I personally feel the compulsion

even though it may be beyond my control. Perkins adds that if we state that a chair

needs to be repaired, we see the need originating in, or being a property of, the chair.

Interestingly, Perkins goes on to note that although these compulsions are seen as

originating within the speaker, they are nevertheless explicitly objective, because they

come from such part of the speaker over which he/she has no control:

(53) Yeah because I need extra cash because my Mom won't give me anyallowance this summer so I need to earn money to be able to go out. (ICE-PHI S1A-022 [B]: 5)

53

In the previous example, speaker B needs to earn money because she is unable to

control his/her mother. In fact, need to may only have an objective interpretation.

Modal need, however, may have both a subjective and an objective interpretation (see

4.9. below). In addition, Perkins (1983, 63) says that the root meaning of need to

denies the involvement of the speaker, but the context in itself may nevertheless have

an illocutionary force of a directive:

(54) You need to ask more close-ended questions to draw out responsesfrom a younger child. (ICE-SIN W2B-016: 42)

Need to also has the past tense form needed to. This is ideal for reporting:

(55) I ran to my car and backed out of the driveway fast, without payingattention to signs or traffic, and I headed toward the highway. I needed togo someplace where I could think, where I could breathe; I felt as if I werechoking. (FROWN P-19: 122–125)

(56) What business furthermore had she to tell an old man that he neededto get rid of the stubble around his chin? (ICE-SIN W2F-020: 87)

4.4.1. Frequency of need to

0

50

100

150

200

250

300

350

400

450


SpokenWritten

Figure 16. Frequency of need to in ICE-PHI, US, ICE-SIN and ICE-GB

54

Quasi-modal need to appears to be frequent in all corpora. In fact, Collins (2005, 259)

says that the verb has “undergone a spectacular increase in recent BrE and AmE”.

This can without a doubt be attributed to the syntactic flexibility of need to, such as

the ability to form different tenses. Interestingly, the spectacular increase in AmE

which Collins refers to has not reached the Philippines: the frequency in ICE-PHI in

spoken language is rather low compared to SBC (173 / 423). The number of need to in

PhiE is also higher in written than in spoken language (200 / 173), which makes the

variety unique. Need to is the first of these verbs that has not caught on in spoken

PhiE, and it also suggests that the expansion of the use of the modals of obligation is

perhaps not as evident in the Philippines as it is in Singapore. This may be due to

what Trudgill (1999, 227) refers to as colonial lag, a delay in the normal development

of linguistic change. According to him, certain features of a language may not be

adopted as quickly as others in colonial situations, mainly because there is no

common peer-group dialect among children. This may result in a time lag before these

features are acquired. Trudgill adds that this lag usually lasts for about one generation.

The frequency of need to is again higher in spoken ICE-SIN than in spoken ICE-

GB (417 / 263). SinE and BrE have adopted need to into written language as well,

while AmE writers seem more reluctant to do so. On the basis of the results viewed so

far, it seems that the Singaporeans are quite innovative in their use of modals and

quasi-modals of obligation and necessity. In the case of need to, however, it could be

argued that the reason may be “Americanisation” at some scale, as the numbers in

spoken ICE-SIN and in SBC and are extremely close to each other (417 / 423). It is

quite likely that AmE has influenced SinE: the political and economic power of the

USA is indisputable, not to mention the pervasive American popular culture

cultivated by the media and the entertainment industry (see also chapter 2.5.2.2).

55

4.4.1.1. Present tense vs. past tense need to

0 %

10 %

20 %

30 %

40 %

50 %

60 %

70 %

80 %

90 %

100 %

ICE-PHI S

poken

ICE-P

HI Writt

en

US Spok

en

US Writt

en

ICE-S

IN Spo

ken

ICE-S

IN W

ritten

ICE-G

B Spoke

n

ICE-G

B Writt

en


Figure 17. Past and present tense need to in ICE-PHI, US, ICE-SIN and ICE-GB

Even though syntactic flexibility is regarded as one of the reasons for the high

frequency of need to, it must be stated that the percentage of past tense needed to

compared to the present tense forms need to and needs to is surprisingly small in all

dialects. Nevertheless, here are a few examples of the past tense use:

(57) And because of the largeness of the area we needed to have a pilotarea of six blocks which is a doable area to demonstrate that the privatesector can come in and be involved in this. (ICE-PHI S2A-041 [A]: 23)

(58) The rent was too high, but they needed only to stay until he foundsomething cheaper. (ICE-SIN W2B-005: 88)

If we compare the numbers of needed to with those of had to above (see Figure 3.

above), it must be acknowledged that past tense needed to is not being used to its full

potential. In addition, the old Englishes, AmE and BrE, seem to use the past tense

form mainly for reporting, as the number is higher in the written registers. Again, the

fact that PhiE and SinE speakers do not necessarily mark verbs for past tense may be

56

the reason (see 4.1.1.2. above). In contrast, PhiE and SinE use the past tense form

relatively little in writing. The fact that the past tense form is hardly used in spoken

SinE is somewhat surprising considering the high frequency of need to in total.

CONVERSATION

0

5

10

15

20

25

30

35

40

45

50

don't need to doesn't needto

didn't need to will need to would need to may need to


Figure 18. Positive and negative verb forms of need to in ICE-PHI, US, ICE-SIN andICE-GB conversation

FICTION

0

5

10

15

20

25

30

don't need to doesn't needto

didn't need to will need to would need to may need to


Figure 19. Positive and negative verb forms of need to in ICE-PHI, US, ICE-SIN andICE-GB fiction

57

In addition to preferring present tense over past tense, the four varieties do not

seem to favour the other verb forms of need to, namely the negative forms and the

forms which are used to refer to future or hypothetical events. The only forms worth

mentioning are the contracted don’t need to in conversation and the hypothetical

would need to in fiction:

(59) I don't need to see other people. (ICE-SIN S1A-041 [B]: 37)

(60) Of course the chaos when the Supreme Being was discovered tied upand concussed on the floor would be indescribable, but surely they wouldneed to be more than just lucky to win much more time out of mere chaos?(ICE-GB W2F-015: 69)

It should also be noted that PhiE appears to be more resilient to these forms than the

other three, probably because of its low overall frequency of need to. The verb forms

are more frequent in AmE, especially in fiction. This is most likely due to its high

overall frequency.

4.4.1.2. Subjects of need to

PersonalICE-PHI pronoun %

Relativepronoun % Other % Total %

Conversation 97 (N=32) 3 (N=1) 0 (N=0) 100 (N=33)Fiction 100 (N=3) 0 (N=0) 0 (N=0) 100 (N=3)


Fiction 83.9 (N=26) 6.4 (N=2) 9.7 (N=3) 100 (N=31ICE-SIN

Conversation 98.3 (N=59) 1.7 (N=1) 0 (N=0) 100 (N=60)Fiction 100 (N=2) 0 (N=0) 0 (N=0) 100 (N=2)

ICE-GBConversation 95.9 (N=47) 0 (N=0) 4.0 (N=2) 100 (N=49)

Fiction 75 (N=6) 12.5 (N=1) 12.5 (N=1) 100 (N=8)

Table 9. Subjects of need to in ICE-PHI, US, ICE-SIN and ICE-GB conversation andfiction

58

With regard to the subjects of need to, personal pronouns are used the most, while the

frequency of nouns as subjects seem to decline compared to the verbs already

discussed. In fact, there were no nouns as subjects in ICE-PHI, and fairly little in the

other corpora as well. Only AmE seems to have variety in its subjects in conversation,

which can be expected considering the high frequency of the verb in US. However,

one must take into consideration the low number of tokens.

4.5. Supposed to

Be supposed to can be paraphrased as be expected to, as in the phrase every man is

supposed to do his duty (Perkins 1983, 85):

(61) You know what I mean you 'll go to like uhm whatever whateverplace and buy like a cellphone or whatever and then you get somethingthat you paid for uhm like it 's supposed to be made in Finland right. (ICE-PHI S1B-035 [B]: 42)

It is also stated in the Longman Grammar of Spoken and Written English (2000, 500)

that be supposed to appears relatively commonly with the progressive aspect. This,

according to the grammar, marks a personal obligation or a likely occurrence, which

is actually in progress at the present moment, or predicted to occur some time in the

future. This was not a very frequent approach in PhiE or in SinE, but there were a few

instances:

(62) No he’s supposed to be giving it to the woman. (ICE-PHI S1B-023[C]: 225)

(63) And they are supposed to be writing in their English mediumuh in exams or whatever projects. (ICE-SIN S1A-071 [B]: 111)

59

4.5.1. Frequency of supposed to

0

50

100

150

200

250

300

350

400


SpokenWritten

Figure 20. Frequency of supposed to in ICE-PHI, US, ICE-SIN and ICE-GB

Based on the study, supposed to belongs to spoken rather than written language,

supposedly because of its personal nature mentioned above (see chapter 4.5.). PhiE

speakers use supposed to most frequently (369), followed by Singaporeans (208). In

fact, both South-East Asian varieties use supposed to more than their colonisers in

both media: the use of the verbs is surprisingly restricted in AmE and BrE, especially

in writing. Whether the Americans and the British regards supposed to as formal

because of the low frequency in speech, or as colloquial because of the same tendency

in writing, is an intriguing question. It is interesting that the South-East Asian

varieties again surpass the old countries in frequency.

60

4.5.1.1. Present tense vs. past tense supposed to

0 %

10 %

20 %

30 %

40 %

50 %

60 %

70 %

80 %

90 %

100 %

ICE-P

HI Conv

ersati

on

ICE -P

HI Fict

ion

US Con

versa

tion

US Fict

ion

ICE-S

IN Conv

ersati

on

ICE-SIN

Fiction

ICE-G

B Con

versa

tion

ICE-G

B Fiction


Figure 21. Present and past tense supposed to in ICE-PHI, US, ICE-SIN and ICE-GBconversation and fiction

Compared to the frequency of past tense have to and need to (see 4.1.1. and 4.4.1.

above), the past tense potential of supposed to is represented more efficiently in

conversation and fiction in all of the varieties. Past tense forms account for over 20%

in all four dialects, apart from SinE fiction, in which there were only present tense

tokens. In AmE fiction, the past tense use accounts for over 75%, which is remarkable

considering the low number of tokens, but not so remarkable when taking into

consideration the reporting nature of fiction. As was the case with past tense had to

and need to, PhiE and SinE users have not adopted past tense supposed to as markedly

as AmE and BrE users.

61

ICE-PHI CONVERSATION

0

50

100

150

200

250

1 ST PERSON SG

2ND PERSON SG

3RD PERSO

N SG

1 ST PERSON PL

2ND PERSON PL

3RD PERSON PL

PASSIVE


Figure 22. Persons of be with supposed to in ICE-PHI conversation

US CONVERSATION

0

20

40

60

80

100

120

140

160

180

1 ST PERSON SG

2ND PERSON SG

3RD PERSON SG

1 ST PERSON PL

2ND PERSON PL

3RD PERSON PL

PASSIV

E


Figure 23. Persons of be with supposed to in US conversation

In conversation, both PhiE and AmE speakers use present tense supposed to

rather than past tense. Third person singular he/she/it is supposed to is the dominant

form used in speech, but past tense he/she/it was supposed to is by no means rare. In

addition, there is a lot more variation in PhiE as to what person, and therefore, what

form of be, is used:

(64) When I graduated actually before I graduated I was supposed to goabroad but the government withheld my passport. (ICE-PHI S1A-

62

009 [B]: 31)

(65) I think they’re supposed to deliberate on it after the end of March.(ICE-PHI S1A-051 [B]: 104)

(66) So not that you’re supposed to make mistakes it’s just that you can'tavoid mistakes. (ICE-PHI:S1A-058 [B]: 182) (passive)

There were only three instances of supposed to in PhiE fiction. One of these was

a 1st person singular present tense and the others were 3rd person singular past tense:

(67) I know I 'm not supposed to go out. (ICE-PHI W2F-015: 7)

(68) By age seventeen, Emilio stopped wondering how his father knew hewas smiling when he was supposed to be unconscious. (ICE-PHI W2F-014: 89)

(69) Vincent, on the other hand, stayed close to the pipe issuing the waterand tried to enjoy the hot bath for what it was supposed to be - - and thewater was indeed very hot, and the weather was particularly warm. (ICE-PHI W2F-020: 9)

US FICTION

0

5

10

15

20

25

30

1 ST PE

RSON SG

2ND PERSON SG

3RD PERSON SG

1 ST PE

RSON PL

2ND PERSON PL

3RD PERSON PL

PASSIV

E


Figure 24. Persons of be with supposed to in US fiction

There is a bit more variation in US fiction in the forms of be, although one must note

that there were a only a few instances in the Frown fiction section (N=16). Past tense

is more frequent in fiction, again most likely because of the reporting nature of the

text type.

63

ICE-SIN CONVERSATION

0

20

40

60

80

100

120

140

160

180

200

1 ST PERSON SG

2ND PERSON SG

3RD PERSON SG

1 ST PERSON PL

2ND PERSON PL

3RD PERSON PL

PASSIVE


Figure 25. Persons of be with supposed to in ICE-SIN conversation

ICE-GB CONVERSATION

0

20

40

60

80

100

120

140

1 ST PERSON SG

2ND PE

RSON SG

3RD PERSO

N SG

1 ST PE

RSON PL

2ND PERSON PL

3RD PERSON PL

PASSIV

E


Figure 26. Persons of be with supposed to in ICE-GB conversation

The present tense use of supposed to is also more frequent in SinE conversation

than past tense, but not overwhelmingly so. It must be acknowledged that past tense

forms of be were found in almost each person, even though Fong (2004, 87) says that

be is optional in this phrase in SinE. There are no remarkable differences between

SinE and BrE: 3rd person singular form is supposed to is most frequent in both dialects

followed by a variety of the other forms. Considering the obliging nature of supposed

64

to, it could be argued that it is odd that the frequency of second person you are

supposed to is so low in all of these dialects. In fact, the passive you are supposed to

was more frequent than the second person form in both varieties, which demonstrates

that impersonalising an obligation is a frequent approach: it is more acceptable to

mask the notion of what someone is expected to do in an utterance which applies to

no one in particular rather than someone who is singled out.

As was the case in ICE-PHI and Frown, there were only a couple instances of

supposed to in ICE-SIN fiction and six in ICE-GB fiction:

(70) “Strangers are not supposed to be walking in and out -" "I'm hisson," he said. (ICE-SIN W2F-004: 7)

(71) The plants are supposed to be green with big healthy leaves. (ICE-SIN W2F-015: 29)

(72) I knew I was supposed to cover for them for a few days, then for afew more days I thought something might have delayed them - a ferrystrike or airport problems or something. (ICE-GB W2F-006: 235)

4.5.1.2. Subjects of supposed to


Conversation 79.2 (N=80) 13.9 (N=14) 6.9 (N=7) 100 (N=101)Fiction 100 (N=3) 0 (N=0) 0 (N=0) 100 (N=3)

USConversation 50 (N=8) 37.5 (N=6) 12.5 (N=2) 100 (N=16)




Fiction 50 (N=2) 25 (N=1) 25 (N=1) 100 (N=4)

Table 10. Subjects of supposed to in ICE-PHI, US, ICE-SIN and ICE-GBconversation and fiction

65

The subjects of supposed to show no difference to the subjects of the other modal

verbs: personal pronouns are the most frequent subject type followed by nouns and

other subjects. In written AmE, however, all subjects were nouns, but this is due to the

low number tokens in the Frown (N=2).

4.6. Have got to

Have got to is semantically quite like have to, which was already discussed in chapter

4.1. Nevertheless, there are differences between these verbs. Palmer (1979, 92) says

that have got to represents a more colloquial style while have to is more formal.

Perkins (1983, 60) also says that have got to is more impersonal, and that it is required

in situations where a certain amount of politeness is required, i.e. in situations where

must would be too intrusive:

(73) Exactly so we’ve got to get our thesis good you know.(ICE-PHI S1A-017 [B]: 178)

In addition, while have to is never subjective, have got to covers a range of meanings

from subjective (74) to more objective (75). However, have got to is never

performative:

(74) Please take that down. You’ve got to read that. (ICE-PHI S1B-018 [A]: 31–32)

(75) But the thing is ya lah only the with the children we alsogot to keep an eye on them and then make sure they don't fall off orwhatever. (ICE-SIN S1A-094 [B]: 211)

Have got to also has an epistemic meaning:

(76) There's got to be more schools around here that have room for aPuppet Lady. You should try to get another gig as soon as this one'sfinished. (FROWN K05: 61–63)

66

4.6.1 Frequency of have got to

0

50

100

150

200

250

300

350

400


SpokenWritten

Figure 27. Frequency of have got to in ICE-PHI, US, ICE-SIN and ICE-GB

Have got to is often reduced to simple got to or gotta. These two reduced forms were

also studied. The results of this study confirm those of the LGSW study above (see

chapter 3.) in that have got to is more frequent in BrE than AmE. This has had an

impact on the language of the colonies: have got to is rare PhiE but frequent in SinE.

The verb is the most frequent in spoken BrE (342), followed by spoken SinE (263),

but virtually non-existent in spoken PhiE (43) and AmE (37). The case of have got to

is also the first instance of BrE having more tokens than SinE. In addition, have got to

appears to be rare in written language in all varieties and non-existent in written PhiE.

This is expected considering the colloquial and impersonal nature of the verb.

67

4.6.1.1. Root and epistemic have got to

0

100

200

300

400

500

600

700

ICE-PHI R

OOT

US ROOT

ICE-S

IN R

OOT

ICE-G

B ROOT

ICE-P

HI EPISTEM

IC

US EPISTEMIC

ICE-S

IN E

PISTEMIC

ICE-G

B EPISTEMIC

ConversationFiction

Figure 28. Frequency of root and epistemic have got toin ICE-PHI, US, ICE-SIN and ICE-GB conversation and fiction

Epistemic have got to is extremely rare, as was epistemic have to. There was only one

instance in both spoken ICE-PHI and ICE-SIN, a few in the Frown, but none in SBC

or ICE-GB, which is surprising taking into account the overall frequency of have got

to in BrE:

(77) It’s got to happen and you know James is like I 'm so excited to goshopping with you baby. (ICE-PHI S1A-015 [A]: 157)

(78) She got to call up one of her boys. (ICE-SIN S1A-069 [A]: 223)

(79) Don't you see, that's got to be the answer? (FROWN K04: 147)

In (77), the speaker is confident that something will soon take place, while the speaker

in (78) is confident that one of her interlocutors, who is speaking on the phone, is

talking to a boyfriend. In example (79), the speaker is simply confident about the

nature of the situation.

68

4.6.1.2. Subjects of have got to


Conversation 100 (N=12) 0 (N=0) 0 (N=0) 100 (N=12)Fiction 0 (N=0) 0 (N=0) 0 (N=0) 0 (N=0)

USConversation 0 (N=0) 100 (N=1) 0 (N=0) 100 (N=1)

Fiction 72.7 (N=16) 0 (N=0) 27.3 (N=6) 100 (N=22)ICE-SIN



Fiction 100 (N=2) 0 (N=0) 0 (N=0) 100 (N=2)

Table 11. Subjects of have got to in ICE-PHI, US, ICE-SIN and ICE-GBconversation and fiction

The subjects of have got to confirm further the notion that personal pronouns are the

most frequent subject of modals of obligation and necessity. Nouns are even more on

the decline and the choice of other subjects seems more random than with the verbs

already discussed. In conclusion, ICE-SIN is the only corpus which shows minor

versatility in the choice of subject with have got to, and again, these other subjects are

mainly instances with no subject at all.

4.7. Ought to

Root ought to expresses weak obligation in both subjective and objective uses,

similarly to root should (Collins 2005, 262). According to Declerk (1992, 377), ought

to can be used to express a non-forceful recommendation (80) or to make a simple

assertion of the convenience of the state of affairs (81)

(80) Washing can be done in a basin of clean water. If so, you ought tochange water several times. (ICE-SIN W2B-032: 99–100)

69

(81) But why are these properties found only in Philippine papaya? Thatought to be the subject of further research. (ICE-PHI W2E-00165–66)

Palmer (1979, 100) says that ought to and should are largely interchangeable. He also

adds that there is often an implication that the event will not, or did not ever take

place, i.e. there is an implication of non-actuality: I ought to be ashamed but I can’t.

According to Coates (1983, 79), epistemic ought to can be paraphrased as ‘meant to’

or ‘supposed to’:

(82) Because I I think there're a lot of mechanics they ought to know righthow to uhm manage the O H T uhm a lot of other classroom mechanicswhich I thought would be helpful to them because now I'm looking back Ifelt find a lot of lecturers and tutors don't really know how to control aclass and how to read the faces of pupils of of the students in in university.(ICE-SIN S1A-076 [B]: 13)

(83) “That pot's been brewing since morning." "Then it ought to beready." (FROWN L-14: 176–177)

Nuyts (2002, 174) in turn says that when uttering ought to, the speaker has weaker

evidence of the situation at hand than with must or with have to, Perkins (1983, 54–55)

adds that ought to is also slightly more polite than must, especially in its negative use:

(84) You oughtn’t to pick your nose in public. (cf. mustn’t)

Surprisingly, there were no examples of oughtn’t in any of the corpora. Positive ought

to is also declining in frequency, as will be demonstrated next.

70

4.7.1 Frequency of ought to

0

50

100

150

200

250


SpokenWritten

Figure 29. Frequency of ought to in ICE-PHI, US, ICE-SIN and ICE-GB

According to the results, ought to has clearly not caught on in either PhiE or SinE. It

is indeed the first modal verb of obligation and necessity which had lower frequency

in both speech and writing in PhiE and SinE. Ought to is nevertheless used in AmE

and BrE, but even then, only scarcely.

It is interesting that the frequency in written AmE is high (201), as many of these

verbs under scrutiny have been scarce in the Frown. In contrast, the number of tokens

in spoken language is higher in BrE (105). The same trend can be noted in PhiE and

SinE in smaller scale. It can then be argued that the use of ought to in PhiE and SinE

has been, at least to some extent, influenced by the language of the colonising

countries, but perhaps the profusion of other similar verbs has contributed to its

disfavour.

71

4.7.1.1. Root and epistemic ought to and subjects of ought to

0

10

20

30

40

50

60

70

80

90

100

ICE-P

HI ROOT

US ROOT

ICE-S

IN ROOT

ICE-G

B ROOT

ICE-PHI E

PISTEMIC

US EPIS

TEMIC

ICE-SIN

EPIS

TEMIC

ICE-G

B EPIS

TEMIC

SpokenWritten

Figure 30. Frequency of root and epistemic ought to in ICE-PHI, US, ICE-SIN andICE-GB conversation and fiction

Root ought to was not found in ICE-PHI private dialogue and creative writing, nor in

ICE-SIN creative writing, which confirms the notion that ought to is not favoured by

PhiE and SinE. Epistemic ought to is also extremely rare: it was not found in ICE-PHI,

spoken US nor ICE-GB, and the only instance found in ICE-SIN conversation was

example 82. However, there were quite a few tokens of epistemic ought to in the

Frown corpus:

(85) "Manj, stay here. Make sure that we question everybody in everyhouse on this block. Neighboring blocks too. We ought to be able to findsomebody who saw something out of place - anything odd.” (FROWN L-4:237–240)

(86) Salina was dark when they passed through. A couple of dogs ran outand challenged them, but they retreated once they'd said their piece. "Theprairie never stops," Thomas said, longing for a soft bed. "We ought tocome into Lincoln about dawn. We'll shake down there." (FROWN N-04:207–213)

72

The Frown corpus was also the only one that had nouns as subjects with ought to

(14.2%, N=2). All other tokens in the corpora had personal pronouns as subjects.

4.8. Had better

By using the verbal expression had better, the speaker gives the further information,

that in his/her view, it ‘is better’ that an event take place, or, as Palmer (1979, 69) puts

it, gives advice and is fairly firm about it. However, this does not entail that the source

of the obligation is necessarily the speaker, and the expression is therefore said to be

objective (Perkins 1983, 63): We had better dress up warmly.

In addition, had better expresses only deontic modality. The use of had better

then entails that something should be brought about, or in contrary, avoided (Perkins

1983, 9; 64). While epistemic meaning relates to knowledge, deontic relates to duties:

(87) And then when I told her that you are uhm you are uh up-to-date onthe literature she said we 'd better give it to Tish. (ICE-PHI S1A-003 [B]: 234)

(88) So, to begin with, the reason for this note/letter/memo is becausePaul came over to the hostel at Eusoff College and told me you called andwould be leaving the ashram where you are now by the end of March andthat I had better write you something before you're again "SOMEWHEREOUT THERE" and where nobody can reach you by mail/air/phone orwhatever the latest technology has invented. (ICE-SIN W1B-008: 2)

73

4.8.1. Frequency of had better

0

10

20

30

40

50

60


SpokenWritten

Figure 31. Frequency of had better in ICE-PHI, US, ICE-SIN and ICE-GB

Had better is definitely on the decline in all of the varieties. Its inflexibility in both

syntactic and semantic aspects has made it a marginal quasi-modal. However, the verb

also appears in PhiE and SinE, and the latter has yet again the highest frequency in

spoken language (48). There were quite a few instances where either the subject

(usually a personal pronoun) or the verb had, or in some cases both, had been omitted.

Therefore, to be able to find all tokens, only better was used as the search word. Here

are a couple examples from ICE-SIN correspondence:

(89) Well better end here before it becomes an autobiography! (ICE-SIN W1B-001: 121)

(90) Better stop myself from 'Fa Lao Sao' (can guess what is this?) -ie letting off my frustrations. (ICE-SIN W1B-002: 28)

There was also an interesting example in ICE-PHI, which incorporates should with

had better:

(91) That I 'm not sure we should better ask him. (ICE-PHI S1A-012 [A]: 142)

74

4.8.1.1. Subjects of had better

USPersonal

pronoun % Noun %Relative

pronoun % Total %Conversation 100 (N=4) 0 (N=0) 0 (N=0) 100 (N=4)

Fiction 92 (N=23) 4 (N=1) 4 (N=1) 100 (N=25)

ICE-SINPersonal

pronoun % None % Total %Conversation 81.3 (N=13) 18.7 (N=3) 100 (N=16)

Fiction 100 (N=4) 0 (N=0) 100 (N=4)

Table 12. Subjects of had better in ICE-PHI, US, ICE-SIN and ICE-GB conversationand fiction

Only the Frown corpus and ICE-SIN show variation in the subjects of had better: all

the instances of the verb in ICE-PHI and ICE-GB had a personal pronoun as subject.

The absence of the subject continues to be a frequent phenomenon in SinE, which was

also demonstrated above in examples 89–90.

4.9. Need

Coates says (1983, 50) that the use of modal need is on the decline, and it seems that

its main function is to provide a negative form (need not or needn’t) which expresses

negation of the modal predication. Christophersen and Sandved (1970, 195) add that

need not expresses absence of obligation while must not expresses a negative

obligation. Need also expresses both root (92) and epistemic necessity (93):

(92) Filipinos need not be told how government can affect their lives. In acorrupt dictatorship, citizens must worry not only about censorship,warrantless arrests and summary executions, but also about spirallingprices, “crony capitalism” and an unhealthy business climate. (ICE-PHIW2E-003: 82–83)

(93) It needn’t be a complicated issue, it can be very straightforward. (= itis not necessarily the case that is complicated, Coates 1983, 50)

Perkins (1983, 62) says that modal need may have both a subjective and an objective

interpretation, while quasi-modal need to may only have an objective one:

75

(94) You needn’t got to the bathroom if you don’t need to.

According to Close (1962, 128), the semantic distinction between need and need to

becomes clear in their negative use:

(95) You need not stay. (= you may go)

(96) You don’t need to stay. (= it is not necessary, there is nothing you cando)

(97) We didn’t need to stay. (= staying was unnecessary, we were free tostay or go)

(98) We needn’t have stayed. (= we stayed unnecessarily)

4.9.1. Frequency of need

0

10

20

30

40

50

60


SpokenWritten

Figure 32. Frequency of need in ICE-PHI, US, ICE-SIN and ICE-GB

Coates mentions (1983, 5) that the use of modal need as opposed to the full verb need

is rare. This notion is confirmed by the results of this study. Need is extremely rare in

all varieties, and practically non-existent in spoken AmE, as there were no tokens in

SBC.

76

4.9.1.1. Need not and needn’t

0 %10 %20 %30 %40 %50 %

60 %70 %80 %90 %

100 %

ICE-PHI S

poken

ICE-PHI W

ritten

US Spo

ken

US Writt

en

ICE-SIN Spok

en

ICE-SIN

Writt

en

ICE-G

B Spoken

ICE-G

B Writt

en

need

need not/needn't

Figure 33. Percentages of positive need and negative need not and needn’t in ICE-PHI, US, ICE-SIN and ICE-GB

The number of positive cases of modal need was notably lower than the number of

negative cases in all corpora: in fact, the positive case was not found in ICE-PHI, but

there were a few positive tokens in the Frown (24.3% N=9). In both spoken ICE-SIN

and ICE-GB, only four instances of need were positive; the rest (N=24 in ICE-SIN

and N=18 in ICE-GB) were negative, and more precisely, mostly cases of the full

form need not, not of the contracted needn’t. This confirms the notion that the main

function of the verb is to provide a negative form. However, the numbers are so small

that the verb can be said to be virtually non-existent in all varieties.

77

4.9.1.2. Subjects of need

USPersonal

pronoun % Noun %Determiner

% Total %Fiction 20 (N=1) 60 (N=3) 20 (N=1) 100 (N=5)

ICE-SINPersonal

pronoun% Total %Conversation 100 (N=3) 100 (N=3)

Fiction 0.0 0

ICE-GBPersonal

pronoun % Noun %Indefinite

pronoun % Total %Conversation 100 (N=1) 0 (N=0) 0 (N=0) 100 (N=1)

Fiction 0 (N=0) 50 (N=1) 50 (N=1) 100 (N=2)

Table 13. Subjects of need in ICE-PHI, US, ICE-SIN and ICE-GB conversation andfiction

Need had low frequency in both conversation and fiction, and therefore the

percentages of subjects offer us no further information on the use of the verb,

especially when there were no tokens in ICE-PHI conversation and fiction. Personal

pronouns were again found the most frequent subject, followed by nouns and other

random cases of subjects.

78

5. Discussion

It is clear the PhiE and SinE should not be considered marginal Englishes that are still

strongly influenced by the language of their colonisers, AmE and BrE. This study

shows that these two relatively new varieties of English are not restricted to following

their donor-languages: Filipinos and Singaporeans are innovative in the use of English,

and they have followed a path of their own in the use of modal and quasi-modal verbs

of obligation and necessity.

5.1. Distribution of the modals and quasi-modals of obligation and necessity

0

1000

2000

3000

4000

5000

6000


SpokenWritten

Figure 34. The frequency of all modal and quasi-modal verbs of obligation andnecessity in ICE-PHI, US, ICE-SIN and ICE-GB

There were tokens of all nine modals in both varieties, and in most cases, the verbs

were more frequent in PhiE and SinE than in AmE and BrE. In fact, it was only in the

case of have got to and ought to that BrE had higher frequency than SinE. In PhiE and

AmE, however, there were a few more differences between verbs: on the whole, the

79

verbs which had the lowest frequencies overall were less likely to be more frequent in

PhiE than in AmE, i.e. need to, have got to, ought to, had better and need.

The subjects of the verbs were mostly personal pronouns, but there were quite a

few cases with nouns, indefinite pronouns, relative pronouns, determiners or

existential there as subject. There was also an interesting trend in SinE, where the

subject had been completely omitted from the clause, i.e. there was a tendency to

imply the subject pronoun rather than state it, because the subject is usually clear from

the context (see section 4.1.1.3.).

0

200

400

600

800

1000

1200

1400

1600

have to should must need to supposedto

have got to ought to had better need


Figure 35. The modals and quasi-modals of obligation and necessity in ICE-PHI, US,ICE-SIN and ICE-GB

In the case of the four varieties, have to, should and must are the most frequent

auxiliaries of obligation. In contrast, need to, supposed to, have got to, ought to, had

better and need are on the decline. Must and should are also more frequent in BrE

than in AmE, as was suggested by the previous studies. Unfortunately, it must also be

noted that the number of all of the verbs was surprisingly low in the Frown corpus. I

80

cannot therefore totally agree with the statement made in the Longman Grammar of

Spoken and Written English (2000, 488), that have to is now also quite common in

written language in AmE: neither my study nor Collins’s 2005 study support this view.

Having stated that, it must be added that have to is still more common in written AmE

than need to, supposed to, have got to, ought to, need and had better.

SHOULD

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000

Private dialogue

Public dialogue

Unscripted monologue

Scripted monologue

Non-professional writing

Correspondence

Academic writing

Non-academic writing

Reportage

Instructional writing

Presuasive writing

Creative writing


Figure 36. The frequency of should in different text types

SUPPOSED TO

0 200 400 600 800 1000 1200 1400 1600

Private dialogue

Public dialogue

Unscripted monologue

Scripted monologue

Non-professional writing

Correspondence

Academic writing

Non-academic writing

Reportage

Instructional writing

Presuasive writing

Creative writing


Figure 37. The Frequency of supposed to in different text types

81

In the different text types of the corpora, private dialogue, instructional writing

and persuasive writing were the most prolific in using the modals. This varied

somewhat between the verbs, but the above examples of the use of should and

supposed to illustrate the main uses. Have to, need to, supposed to and got to were

mainly used in speech while should, must, ought to, had better and need were more or

less used in writing.

Epistemic meaning in conversation

ICE-PHI

US

ICE-SIN

ICE-GB

Figure 38. Epistemic meaning in ICE-PHI, US, ICE-SIN and ICE-GB conversation

Epistemic meaning in fiction

ICE-PHI

US

ICE-SIN

ICE-GB

Figure 39. Epistemic meaning in ICE-PHI, US, ICE-SIN and ICE-GB fiction

82

Epistemic meaning was without a doubt secondary to root meaning in all

varieties. There were only a few examples of epistemic use in the cases of most verbs,

with the exception of must, which had a quite high frequency of epistemic use. In

conversation, PhiE speakers are less likely to use epistemic meaning than AmE

speakers, while in contrast, PhiE fiction writers are more likely to do so than AmE

writers. SinE users are also less likely to lean towards epistemic use than BrE users in

both media, although the difference is rather small in conversation (415 / 420). In

conclusion, SinE and BrE users are more accustomed to epistemic meaning than PhiE

and AmE users.

5.2. Cultural aspects

Why is it then that the countries which were colonised by English-speaking people not

more than two hundred years ago have overtaken the original Englishes in the use of

these verbs? Especially in the case of Singapore, why is it that the numbers are higher?

English-medium education is hardly a reason for the figures in ICE-SIN, as English is

used in education in the Philippines as well.

It is generally assumed that western English-speaking people are extremely

aware of their face in conversation, and do not wish to be seen as overpowering.

There is a strive for consensus even though there might be a hierarchical relationship

between the speakers. This is reflected on the use of verbs of obligation, as, for

example, the use of must, which is regarded to express strong obligation, has declined

in frequency. In contrast, eastern philosophies are clear in their insistence on respect

towards those who are more mature, and in the emphasis on an extremely hierarchical

social system. It can then be suggested that the number of verbs of obligation is higher

because it may be that it is more acceptable to oblige people in the Philippines and

Singapore. This is not to say that Asian English speakers wish in any respect to seem

83

overpowering. It is to acknowledge the hierarchical and collective nature of these

cultures, which may be a part of the reason for the high frequency of the verbs of

obligation and necessity.

This view entails that the hierarchical system is quite visible, or rather audible, in

the way people interact. In fact, Singapore did not achieve its position among the most

successful nations in the world by giving its citizens the freedom to act according to

their own wishes and desires. It seems that in Singapore, everyone goes along with

what those above them command and they will continue to do so as long as their

welfare remains secure. Foley (1988, 3) adds that while some Anglophone colonies

have moved from English as a foreign language to English as an L2 and back to

English as a foreign language, English in Singapore is moving from an L2 to a native

language. This is also spurred by the fact that although it has a colonial origin, English

was politically neutral in Singapore, and could therefore promote national unity. This

has without a doubt had an effect in moving SinE farther away from BrE.

According to Bernardo (2004, 29), English in the Philippines is now mainly a

language of access to education, and its function has not been nearly as unifying as it

has in Singapore. Having stated that, the difference between PhiE and AmE almost as

large as that between SinE and BrE. It can be argued therefore, that the difference

could be due to a more or less conscience differentiation from AmE.

It can naturally be argued that computer corpora are not valid to make these sorts

of generalisations, but it is quite evident that corpora provide an objectively-based

method for the study of such social phenomena. For example, Leech and Fallon (1992,

44) examined the LOB and the Brown Corpus and came to the conclusion that

because of the high frequency of words such as perhaps, rather and fairly in LOB,

BrE speakers come across as lacking firmness and decisiveness, which conforms well

84

with the British stereotype. They also say (1992, 45) that the method “claims to be

moving towards a true picture of cultural contrasts, on the grounds that the evidence is

in the corpora, and no other explanation can be found for it”.

85

6. Final Remarks

The Philippines and Singapore are, without a doubt, fruitful areas for any study. The

encounter of the west and the east is unique in both countries. In the Philippine

archipelago, the native cultures of the small islands, and the way of living as one with

the ocean meet the hectic and urban English-speaking environment of the capital city

Manila. In Singapore, the Asian discipline and strive for excellence come together

with western capitalism and English-driven internationalism. These countries were

American and British colonies not more than sixty years ago, but it is nevertheless

evident that Filipinos and Singaporeans have succeeded in making English their own

language and a part of their own culture without having to rigidly model its structure

and use according to AmE and BrE.

This study focused on the expression of the modality of obligation and necessity

in PhiE and SinE. This semantic category, which refers to the speaker’s attitude or

opinion on a certain situation, is realised in English primarily with modal verbs such

as must and should and quasi-modal verbs such as have to and need to. Five electronic

corpora were used to retrieve information on how modals must, should and need, and

quasi-modals have to, have got to, need to, supposed to, ought to and had better are

used in the above varieties. These corpora included the ICE components of PhiE, SinE

and BrE, the SBC and the Frown corpus. The main agenda was to find out to what

extent these new Englishes have retained features from AmE and BrE in the use of

these modal expressions.

There were plenty of tokens of most of the verbs in all corpora, which proves

that the expression of modality of obligation and necessity is a central part of the

English language on all continents. Have to was the most frequent verb in all varieties

followed by should, must, need to, supposed to and have got to respectively. Ought to,

86

had better and need were rare in all varieties. The popularity of have to and should

was attributed to their unthreatening and democratic nature in speech and writing,

while the unpopularity of must was generally thought to be due to its strict and

authoritative nature. These results are well comparable with those of The Longman

Grammar of Spoken and Written English (2005) and Collins (2005). Both of these

studies suggested that have to, should and must are the most frequent verbs used in

this respect.

It also became evident that South-East Asian PhiE and SinE were more prolific

in the use of the modals and quasi-modals of obligation and necessity in both speech

and writing. It was only in the case of have got to and ought to that AmE and BrE had

more tokens than PhiE and SinE. In addition, need to, had better and need were less

frequent in PhiE than in AmE. However, the total numbers of have got to, ought to,

had better and need were marginal and the differences minuscule. Another point made

concerns the frequency of epistemic modality, which refers to the speaker’s

assessment of the situation and evaluation of chances. Overall, epistemic modality

was rare with most verbs with the exception of must. In addition, epistemic modality

was more frequent in SinE and BrE than in PhiE and AmE.

It was suggested that the reason for the trend of favouring the modals and quasi-

modals in the South-East Asian varieties may lie at least partly in the culture of the

Philippines and Singapore. While western English speakers are regarded as striving

for consensus and wishing not to seem overpowering, eastern speakers may be more

accustomed to the expression of obligation and authority. This is not to say that

authority in the east is gained by overpowering one’s interlocutor. Nevertheless, one

must acknowledge the collective nature of eastern societies, which do not function

87

without hierarchy. It can then be argued that hierarchy does not function without

obligation.

It is evident that PhiE and SinE enrich the knowledge we have of the Englishes

of the world. The abundance of phonological, syntactic, semantic and sociolinguistic

features which are yet to be discovered and accounted for in these two languages offer

a wide scope for any linguist. On the basis of this study, the future holds a possible

study in the sociolinguistic field. It would be rewarding to find out to what extent the

reasons for the high frequency of the modals and quasi-modals of obligation and

necessity are connected with Asian philosophies, the way of living, and the way of

situating oneself in society. Another study could take into consideration the other

modal verbs, such as can, may and will. Meanwhile, English may well be the only

language in the world without genuine frontiers, but as long as man-made frontiers

exist, PhiE and SinE will continue to develop in their own directions.

88

References

Corpora

The Freiburg - Brown Corpus of American English. CD-ROM. 1999. EnglischesSeminar, Albert-Ludwigs-Universität, Freiburg.

The International Corpus of English: Great Britain. CD-ROM. 1998. The Survey ofEnglish Usage, University College London, London.

The International Corpus of English: the Philippine Corpus. CD-ROM. 2004.Department of English and Applied linguistics, De La Salle University,Manila.

The International Corpus of English: the Singapore Corpus. CD-ROM. 2002.Department of English Language & Literature, The National University ofSingapore, Singapore.

The Santa Barbara Corpus of Spoken American English Parts I-IV. CD ROM. 2000–2005. Santa Barbara Center for the Study of Discourse, University of California,Santa Barbara.

Literature

Bernardo, Allan B. 2004. “McKinley's questionable bequest: over 100 years ofEnglish in Philippine education.” World Englishes 23:1, 17–31.

Bhat, D.N.S. 1999. The Prominence of Tense, Aspect and Mood. Amsterdam: JohnBenjamins Publishing.

Bolton, Kingsley and Ma Lourdes S. Bautista. 2004. “Philippine English: tensions andtransitions.” World Englishes 23:1, 1–5.

Chafe, Wallace L., John W. DuBois and Sandra A. Thompson. 1991. “Towards a newcorpus of spoken American English.” In English corpus linguistics. Studies inhonour of Jan Svartvik, eds. Karin Aijmer and Bengt Altenberg, 64–82.London: Longman.

Christophersen, Paul and Arthur O. Sandved. 1970. An Advanced English Grammar.Basingstoke: Macmillan.

Chua, S.C. 1964. Report on the Census of Population 1957. Singapore: State ofSingapore.

Close, R.A. 1962. English as a Foreign Language. London: George Allen & Unwin.Coates, Jennifer. 1983. The Semantics of the Modal Auxiliaries. London: Croom Helm.Collins, Peter C. 2005. “The modals and quasi-modals of necessity and obligation in

Australian English and other Englishes.” English World Wide 26:3, 249–273.Davis, Polly. 1977. English structure in focus. Rowley: Newbury House.Dayag, Danilo T. 2004. “The English-language media in the Philippines.” World

Englishes 23:1, 33–45Declerk, Renaat. 1992. A Comprehensive Descriptive Grammar of English. Tokyo:

Kaitakusha.Dixon, L Quentin. 2005. “Bilingual Education Policy in Singapore: An Analysis of Its

Sociohistorical Roots and Current Academic Outcomes.” InternationalJournal of Bilingual Education and Bilingualism 8:1, 25–47.

Fong, Vivienne. 2004. “The verbal cluster.” In Singapore English. A grammaticaldescription, ed. Lisa Lim, 75–104. Amsterdam: John Benjamins Publishing.

Foley, Joseph. 1988. “Studies in Singapore English: Looking Back and LookingForward.” In New Englishes: The Case of Singapore, ed. Joseph Foley, 1–26.

89

Singapore: Singapore University Press.van Gelderen, Elly. 2003. “Aspect in English Modal Complements.” Studia

Linguistica 57:1, 27–43.Gonzalez, Andrew. 1991. “The English of the Philippine print media.” In English

Around the World: Sociolinguistic perspectives, ed. Jenny Cheshire, 333–363.Cambridge: Cambridge University Press.

Gonzalez, Andrew. 2004. “The social dimensions of Philippine English.” WorldEnglishes 23:1, 7–16.

Greenbaum, Sidney. 1990. “Standard English and the international corpus ofEnglish.” World Englishes 9:1, 79–83.

Greenbaum, Sidney. 1991a. “ICE: the International Corpus of English.” EnglishToday 28, 3–7.

Greenbaum, Sidney. 1991b. “The development of the International Corpus ofEnglish.” In English Corpus Linguistics. Studies in Honour of Jan Svartvik,eds. Karin Aijmer and Bengt Altenberg, 83–91. London: Longman.

Greenbaum, Sidney. 1996. “Introducing ICE.” In Comparing English Worldwide: TheInternational Corpus of English, ed. Sidney Greenbaum, 3–12. Oxford:Clarendon Press.

Hall, Robert A. Jr. 1974. Pidgin and Creole Languages. Ithaca: Cornell UniversityPress.

Ho, Mian Lian and John T. Platt. 1993. Dynamics of a contact continuumSingaporean English. Oxford: Clarendon Press.

Huddleston, Rodney. 1988. English Grammar: An Outline. Cambridge: CambridgeUniversity Press.

Huddleston, Rodney and Geoffery K. Pullum 2002. Cambridge Grammar of theEnglish Language. Cambridge: Cambridge University Press.

Kuo, Eddie C.Y. 1978. “Multilingualism and Mass Media Communications inSingapore.” Asian Survey 18:10, 1067–1083.

Lee, K.Y. 2000. From Third World to First: The Singapore Story 1965–2000. NewYork: Harper Collins.

Leech, Geoffery and Roger Fallon. 1992. “Computer corpora – What do they tell usabout culture?” ICAME Journal 16, 29–50.

Ledesma, Heloise Marie L. 2005. “Patterns of Language Preference Among Bilingual(Filipino-English) Boys.” International Journal of Bilingual Education andBilingualism 8:1, 62–80.

Lim, Lisa and Joseph A. Foley. 2004. “English in Singapore and Singapore English.Background and methodology.” In Singapore English. A grammaticaldescription, ed. Lisa Lim, 1–18. Amsterdam: John Benjamins Publishing.

Longman Grammar of Spoken and Written English, 3rd impression. 2000. ed. DouglasBiber. London: Longman.

Lyons, John. 1977. Semantics. Cambridge: Cambridge University Press.MacDougall, John A. and Chew Sock Foon. 1976. “English Language Competence

and Occupational Mobility in Singapore” Pacific Affairs 49:2, 294–312.Martin, J.R. 2004. “Metafunctional profile of the grammar of Tagalog.” In Language

Typology. A functional perspective, ed. Alice Caffarel, 255–304. Amsterdam:John Benjamins Publishing.

McFarland, Curtis D. 2004. “The Philippine language situation.” World Englishes23:1, 59–75.

Myhill, John. 1995. “Change and Continuity in the functions of the American Englishmodals.” Linguistics 33:2, 157–211.

90

Neher, Clarke D. 1999 “The Case for Singapore.” In Singapore Puzzle, ed. MichaelHaas, 39–54. Westport: Praeger Publishers.

Nelson, Gerald. 1996. “The Design of the Corpus.” In Comparing English Worldwide:The International Corpus of English, ed. Sidney Greenbaum, 27–35. Oxford:Clarendon Press.

Nelson, Gerald. 2001. English: An Essential grammar. London: Routledge.Nuyts, Jan. 2002. Epistemic Modality, Language and Conceptualization: a Cognitive-

Pragmatic Perspective. Amsterdam: John Benjamins Publishing.Otanes Fe T. and Bonifacio P. Sibayan. 1969. Language Policy Survey of the

Philippines. Manila: Language Study Center, Philippine Normal College.Palmer, F.R. 1968. A Linguistic Study of the English Verb. London: Longman.Palmer, F.R. 1979. Modality and the English Modals. London: Longman.Palmer, F.R. 1986. Mood and Modality. Cambridge: Cambridge University Press.Papafragou, Anna. 1998. “Inference & word meaning: The case of modal

auxiliaries.” Lingua: International Review of General Linguistics 105:1, 1–47.Papafragou, Anna. 2002. “Modality and theory of mind.” In Modality and its

Interaction with the Verbal System, ed. Sjef Barbiers, 185–204. Amsterdam:John Benjamins Publishing.

Perkins. Michael R. 1983. Modal expressions in English. London: Frances Pinter.Platt, John. 1991. “Social and linguistic constraints on variation in the use of two

grammatical variables in Singapore English.” In English Around the World:Sociolinguistic perspectives, ed. Jenny Cheshire, 376–387. Cambridge:Cambridge University Press.

Platt, John, Heidi Weber and Ho Mian Lian. 1984. The New Englishes. London:Routledge and Kegan Paul.

Quirk, Randolph and Sidney Greenbaum. 1975. University Grammar of English.London: Longman.

Quirk, Randolph, Sidney Greenbaum, Geoffery Leech and Jan Svartvik. 1985. AComprehensive Grammar of the English Language. London: Longman.

SarDesai, D.R. 1997. Southeast Asia: Past & Present, 4th edition. Boulder: WestviewPress.

Tinio, Rolando S. 1990. A Matter of Language: Where English Fails. Diliman,Quezon City. University of the Philippines Press.

Tollefson, James W. 1991. Planning Language, Planning Inequality. Language policyin the community. London: Longman.

Trudgill, Peter. 1999. “A window on the past: ‘Colonial lag’ and New Zealandevidence for the phonology of nineteenth century English.” American Speech74:3, 227–239.

Tupas, T. Ruanni F. 2004. “The politics of Philippine English: neocolonialism, globalpolitics, and the problem of postcolonialism.” World Englishes 23:1, 47–58.

Turnbull, C.M. 1977. A History of Singapore 1819–1975. Oxford: Oxford UniversityPress.

Warner, Anthony R. 1993. English Auxiliaries Structure & History. Cambridge:Cambridge University Press.

Wells, J.C. 1982. Accents of English, Vol 3. Cambridge: Cambridge University Press.Woisetschlager, Erich F. 1985. Semantic Theory of the English Auxiliary System. New

York: Garland Publishing.von Wright, Georg Henrik. 1999. “Deontic Logic: A personal View.” Ratio Juris 12:1,

26–38.Zhiming, Bao. 2001. “The origins of empty categories in Singapore English.” Journal

91

of Pidgin and Creole Languages 16:2, 275–319.Ziegeler, Debra. 2000. Hypothetical modality. Grammaticalization of an L2 dialect.

Amsterdam: John Benjamins Publishing.

Electronic sources

Frown manual: Manual of Information to accompany The Freiburg - Brown Corpusof American English. CD-ROM. Hundt, Marianne, Andrea Sand and PaulSkandera. 1999. Englisches Seminar, Albert-Ludwigs-Universität, Freiburg.

ICE-GB manual: The International Corpus of English: Great Britain. CD-ROM.1998. The Survey of English Usage, University College London, London.

ICE-PHI manual: International Corpus of English: the Philippines Corpus. CD-ROM.2004. Department of English and Applied Linguistics, De La Salle University,Manila.

ICE-SIN manual: International Corpus of English: the Singapore Corpus UserManual. CD-ROM. Nelson, Gerald. 2002. Department of English Language &Literature, The National University of Singapore, Singapore.

Relief Web. 1996. United Nations Office for the Coordination of HumanitarianAffairs. 17 Apr 2006. >http://www.reliefweb.int/mapc/asi_se/reg/seasia.html<

Singapore Department of Statistics. 2000. Singapore census of population, 2000,advance data release n:o 3: Literacy and language.>http://www.singstat.gov.sg/papers/c2000/adr-literacy.pdf<

SBC Manual: The Santa Barbara Corpus of Spoken American English Parts I-IV. CDROM. 2000–2005. Santa Barbara Center for the Study of Discourse,University of California, Santa Barbara.

http://www.reliefweb.int/mapc/asi_se/reg/seasia.html

http://www.singstat.gov.sg/papers/c2000/adr-literacy.pdf

92

Appendix

All frequencies have been normalised to tokens per one million words.

Number of words in each corpora:

Total Spoken Written Conversation FictionICE-PHI 956 000 556 000 400 000 182 000 40 000

US 1 266 972 266 972 1 000 000 69 557 233 493ICE-SIN,ICE-GB 1 000 000 600 000 400 000 200 000 40 000

Frequencies for Figure 1. Modals and quasi-modals of necessity in Australian, British,New Zealand and American English (figures normalised to tokens per one millionwords). (Collins 2005, 269)

ICE-AUS ICE-GB ICE-NZ C-US TOTALmust 613 675 714 402 2404

have to 1311 1244 1182 1384 5121should 1141 1065 1577 850 4633need to 343 208 338 473 1362

have got to 332 339 229 173 1073ought to 36 80 58 51 225

need 19 26 20 15 80had better 48 33 36 41 158

may/might aswell 14 15 22 5 56

Total 3857 3685 4176 3394 15112

Frequencies for Figure 2. Frequency of have to in ICE-PHI, US, ICE-SIN and ICE-GB (figures normalised to tokens per one million words, raw figures in parentheses).

have to ICE-PHI US ICE-SIN ICE-GBSpoken 1709 (950) 1768 (472) 1776 (1066) 1310 (786)Written 808 (323) 582 (582) 955 (382) 910 (364)

93

Frequencies for Figure 3. Frequency of root and epistemic have to in ICE-PHI, US,ICE-SIN and ICE-GB conversation and fiction (figures normalised to tokens per onemillion words, raw figures in parentheses).

ICE-PHI US ICE-SIN ICE-GBhave to Root Root Root Root

Conversation 2418 (440) 1826 (127) 1760 (352) 1390 (278)Fiction 1075 (43) 917 (214) 1100 (44) 1100 (44)

ICE-PHI US ICE-SIN ICE-GBEpistemic Epistemic Epistemic Epistemic


Frequencies for Figure 4. Present tense have to and has to vs. past tense had to inICE-PHI, US, ICE-SIN and ICE-GB.

ICE-PHI Present tense % Past tense % Total %Spoken 85.5 (812) 14.5 (138) 100 (950)Written 75 (243) 25 (80) 100 (323)

US Present tense % Past tense % Total %Spoken 75 (354) 25 (118) 100 (471)Written 62.4 (363) 37.6 (219) 100 (582)

ICE-SIN Present tense % Past tense % Total %Spoken 91.4 (974) 8.6 (92) 100 (1066)Written 70.5 (269) 29.5 (113) 100 (382)ICE-GB Present tense % Past tense % Total %Spoken 80 (629) 20 (157) 100 (786)Written 78.8 (287) 21.2 (77) 100 (364)

Frequencies for Figure 5. Positive and negative verb forms of have to in ICE-PHI, US,ICE-SIN and ICE-GB conversation (figures normalised to tokens per one millionwords, raw figures in parentheses).

Conversation ICE-PHI US ICE-SIN ICE-GBdo not have to 5 (1) 0 (0) 0 (0) 0 (0)

don't have to 99 (18) 72 (5) 130 (26) 100 (20)doesn't have to 5 (1) 43 (3) 25 (5) 10 (2)did not have to 5 (1) 0 (0) 0 (0) 0 (0)

didn't have to 33 (6) 43 (3) 5 (1) 10 (2)have had to 0 (0) 0 (0) 0 (0) 10 (2)has had to 0 (0) 0 (0) 0 (0) 10 (2)

will have to 17 (3) 58 (4) 70 (14) 110 (22)would have to 0 (0) 58 (4) 0 (0) 0 (0)won't have to 0 (0) 0 (0) 5 (1) 5 (1)

94

may have to 0 (0) 0 (0) 25 (5) 5 (1)might have to 0 (0) 43 (3) 5 (1) 15 (3)

Frequencies for Figure 6. Positive and negative verb forms of have to in ICE-PHI, US,ICE-SIN and ICE-GB fiction (figures normalised to tokens per one million words,raw figures in parentheses).

Fiction ICE-PHI US ICE-SIN ICE-GBdon't have to 50 (2) 21 (5) 50 (2) 0 (0)

doesn't have to 25 (1) 0 (0) 0 (0) 0 (0)did not have to 0 (0) 9 (2) 0 (0) 0 (0)

didn't have to 0 (0) 0 (0) 25 (1) 0 (0)have had to 0 (0) 0 (0) 25 (1) 0 (0)had had to 0 (0) 13 (3) 25 (1) 25 (1)

will have to 75 (3) 77 (18) 0 (0) 75 (3)would have to 100 (4) 94 (22) 25 (1) 225 (9)

will have had to 0 (0) 9 (2) 0 (0) 0 (0)would have had

to 0 (0) 4 (1) 25 (1) 0 (0)won't have to 0 (0) 9 (2) 0 (0) 0 (0)wouldn't have

to 0 (0) 13 (3) 0 (0) 0 (0)may have to 0 (0) 0 (0) 0 (0) 0 (0)

might have to 25 (1) 13 (3) 0 (0) 0 (0)

Frequencies for Figure 7. Subject types of have to ICE-PHI, US, ICE-SIN and ICE-GB conversation and fiction.

Personalpronoun % Noun %

Indefinitepronoun %

Relativepronoun %

ICE-PHI 96.3 (468) 2.5 (12) 0.2 (1) 0.6 (3)US 81.0 (285) 13.5 (47) 2.6 (9) 1.1 (4)

ICE-SIN 92.5 (368) 3.2 (13) 0.5 (2) 1.0 (4)ICE-GB 94.2 (308) 3.9 (13) 0.3 (1) 0.3 (1)

None % Determiner %Existential

there % Total %ICE-PHI 0 (0) 0.4 (2) 0 (0) 100 (486)

US 0 (0) 1.4 (5) 0.6 (2) 100 (352)ICE-SIN 2.3 (9) 0.5 (2) 0 (0) 100 (398)ICE-GB 0.9 (3) 0 (0) 0.3 (1) 100 (327)

95

Frequencies for Figure 8. Frequency of should in ICE-PHI, US, ICE-SIN and ICE-GB(figures normalised to tokens per one million words, raw figures in parentheses).

should ICE-PHI US ICE-SIN ICE-GBSpoken 1248 (694) 588 (157) 1482 (889) 945 (567)Written 1195 (478) 788 (788) 1395 (558) 1295 (518)

Frequencies for Figure 9. Frequency of should in ICE-PHI, US, ICE-SIN and ICE-GBconversation and fiction (figures normalised to tokens per one million words, rawfigures in parentheses).

ICE-PHI US ICE-SIN ICE-GBshould Root Root Root Root


ICE-PHI US ICE-SIN ICE-GBshould Epistemic Epistemic Epistemic Epistemic


ICE-PHI US ICE-SIN ICE-GBshould Subjunctive Subjunctive Subjunctive Subjunctive


ICE-PHI US ICE-SIN ICE-GBshould Would Would Would Would


Frequencies for Figure 10. Should have and should’ve in ICE-PHI, US, ICE-SIN andICE-GB conversation and fiction (figures normalised to tokens per one million words,raw figures in parentheses).

should have /should've ICE-PHI US ICE-SIN ICE-GB


96

Frequencies for Figure 11. Should not and shouldn’t in ICE-PHI, US, ICE-SIN andICE-GB.

should not % shouldn't % Total %ICE-PHI 30.4 (7) 69.6 (16) 100 (23)

US 8.3 (2) 91.6 (22) 100 (24)ICE-SIN 28.6 (10) 71.4 (25) 100 (35)ICE-GB 3.6 (1) 96.4 (27) 100 (28)

Frequencies for Figure 12. Percentages of subject types of should in ICE-PHI, US,ICE-SIN and ICE-GB.

Personalpronoun % Noun % None % Determiner %

ICE-PHI 86.0 (208) 10.3 (25) 0.4 (1) 1.2 (3)US 82.5 (142) 11.0 (19) 1.7 (3) 1.2 (2)

ICE-SIN 81.4 (245) 7.0 (21) 8.3 (25) 2.3 (7)ICE-GB 81.4 (162) 5.5 (11) 4.5 (9) 5.0 (10)

Relativepronoun %

Existentialthere %

Indefinitepronoun % Total %

ICE-PHI 1.7 (4) 0.4 (1) 0 (0) 100 (242)US 1.2 (2) 1.2 (2) 1.2 (2) 100 (172)

ICE-SIN 0.7 (2) 0.3 (1) 0 (0) 100 (301)ICE-GB 1.0 (2) 2.0 (4) 0.5 (1) 100 (199)

Frequencies for Figure 13. Frequency of must in ICE-PHI, US, ICE-SIN and ICE-GB(figures normalised to tokens per one million words, raw figures in parentheses).

must ICE-PHI US ICE-SIN ICE-GBSpoken 568 (316) 187 (50) 993 (596) 507 (304)Written 813 (325) 668 (668) 1068 (427) 915 (366)

Frequencies for Figure 14. Root and epistemic must in ICE-PHI, US, ICE-SIN andICE-GB conversation and fiction (figures normalised to tokens per one million words,raw figures in parentheses).

ICE-PHI US ICE-SIN ICE-GBmust Root Root Root Root


97

ICE-PHI US ICE-SIN ICE-GBmust Epistemic Epistemic Epistemic Epistemic


Frequencies for Figure 15. Must have and must’ve in ICE-PHI, US, ICE-SIN andICE-GB conversation and fiction (figures normalised to tokens per one million words,raw figures in parentheses).

must have/must've ICE-PHI US ICE-SIN ICE-GB


Frequencies for Figure 16. Frequency of need to in ICE-PHI, US, ICE-SIN and ICE-GB (figures normalised to tokens per one million words, raw figures in parentheses).

need to ICE-PHI US ICE-SIN ICE-GBSpoken 173 (96) 423 (113) 417 (250) 263 (158)Written 200 (80) 159 (159) 285 (114) 253 (101)

Frequencies for Figure 17. Past and present tense need to in ICE-PHI, US, ICE-SINand ICE-GB.

ICE-PHI Present tense % Past tense % Total %Spoken 92 (92) 8 (8) 100 (100)Written 95 (76) 5 (4) 100 (80)

US Present tense % Past tense % Total %Spoken 92.0 (104) 8.0 (9) 100 (113)Written 85.4 (135) 14.6 (23) 100 (158)

ICE-SIN Present tense % Past tense % Total %Spoken 99.2 (248) 0.8 (2) 100 (250)Written 94.7 (108) 5.3 (6) 100 (114)ICE-GB Present tense % Past tense % Total %Spoken 93.0 (147) 7.0 (11) 100 (158)Written 90.1 (91) 9.9 (10) 100 (101)

98

Frequencies for Figure 18. Positive and negative forms of need to in ICE-PHI, US,ICE-SIN and ICE-GB conversation (figures normalised to tokens per one millionwords, raw figures in parentheses).

Conversation ICE-PHI US ICE-SIN ICE-GBdon't need to 11 (2) 43 (3) 30 (6) 30 (6)

doesn't need to 5 (1) 0 (0) 0 (0) 0 (0)didn't need to 5 (1) 14 (1) 0 (0) 0 (0)

will need to 0 (0) 0 (0) 0 (0) 5 (1)would need to 0 (0) 14 (1) 0 (0) 10 (2)

may need to 0 (0) 0 (0) 5 (1) 0 (0)

Frequencies for Figure 19. Positive and negative forms of need to in ICE-PHI, US,ICE-SIN and ICE-GB fiction (figures normalised to tokens per one million words,raw figures in parentheses).

Fiction ICE-PHI US ICE-SIN ICE-GBdon't need to 0 (0) 9 (2) 0 (0) 0 (0)

doesn't need to 0 (0) 9 (2) 0 (0) 0 (0)didn't need to 0 (0) 13 (3) 0 (0) 0 (0)

will need to 0 (0) 9 (2) 0 (0) 0 (0)would need to 0 (0) 4 (1) 25 (1) 25 (1)

may need to 0 (0) 4 (1) 0 (0) 0 (0)

Frequencies for Figure 20. Frequency of supposed to in ICE-PHI, US, ICE-SIN andICE-GB (figures normalised to tokens per one million words, raw figures inparentheses).

supposed to ICE-PHI US ICE-SIN ICE-GBSpoken 369 (205) 165 (44) 208 (125) 137 (82)Written 88 (35) 55 (55) 53 (21) 43 (17)

Frequencies for Figure 21. Present and past tense supposed to in ICE-PHI, US, ICE-SIN and ICE-GB conversation and fiction.

ICE-PHI Present tense % Past tense % Total %Conversation 75.2 (76) 24.8 (25) 100 (101)

Fiction 66.7 (2) 33.3 (1) 100 (3)US Present tense % Past tense % Total %

Conversation 62.5 (10) 37.5 (6) 100 (16)Fiction 33.3 (5) 66.7 (10) 100 (15)

ICE-SIN Present tense % Past tense % Total %Conversation 80 (72) 20 (18) 100 (90)

99

Fiction 100 (2) 0 (0) 100 (2)ICE-GB Present tense % Past tense % Total %

Conversation 73.7 (42) 26.3 (15) 100 (57)Fiction 50 (2) 50 (2) 100 (4)

Frequencies for Figure 22. Persons of be with supposed to in ICE-PHI conversation(figures normalised to tokens per one million words, raw figures in parentheses).

ICE-PHIPresent tenseconversation

Past tenseconversation

1st person sg 55 (10) 44 (8)2nd person sg 44 (8) 5 (1)3rd person sg 137 (25) 71 (13)1st person pl 82 (15) 11 (2)

2nd person pl 5 (1) 0 (0)3rd person pl 66 (12) 5 (1)

passive 27 (5) 0 (0)

Frequencies for Figure 23. Persons of be with supposed to in US conversation (figuresnormalised to tokens per one million words, raw figures in parentheses).

USPresent tenseconversation




passive 0 (0) 0 (0)

Frequencies for Figure 24. Persons of be with supposed to in US fiction (figuresnormalised to tokens per one million words, raw figures in parentheses).

US Present tense fiction Past tense fiction1st person sg 0 (0) 0 (0)

2nd person sg 4 (1) 0 (0)3rd person sg 4 (1) 20 (5)1st person pl 4 (1) 4 (1)


passive 4 (1) 0 (0)

100

Frequencies for Figure 25. Persons of be with supposed to in ICE-SIN conversation(figures normalised to tokens per one million words, raw figures in parentheses).

ICE-SINPresent tenseconversation


1st person sg 50 (10) 20 (4)2nd person sg 25 (5) 5 (1)3rd person sg 155 (31) 35 (7)

1st person pl 55 (11) 5 (1)2nd person pl 0 (0) 0 (0)3rd person pl 50 (10) 10 (2)

passive 25 (5) 15 (3)

Frequencies for Figure 26. Persons of be with supposed to in ICE-GB conversation(figures normalised to tokens per one million words, raw figures in parentheses).

ICE-GBPresent tenseconversation




passive 15 (3) 5 (1)

Frequencies for Figure 27. Frequency of have got to in ICE-PHI, US, ICE-SIN andICE-GB (figures normalised to tokens per one million words, raw figures inparentheses).

have got to ICE-PHI US ICE-SIN ICE-GBSpoken 43 (24) 37 (10) 263 (158) 342 (205)Written 2,5 (1) 41 (41) 40 (16) 30 (12)

101

Frequencies for Figure 28. Frequency of root and epistemic have got to in ICE-PHI,US, ICE-SIN and ICE-GB conversation and fiction (figures normalised to tokens perone million words, raw figures in parentheses).

ICE-PHI US ICE-SIN ICE-GBhave got to Root Root Root Root




Frequencies for Figure 29. Frequency of ought to in ICE-PHI, US, ICE-SIN and ICE-GB (figures normalised to tokens per one million words, raw figures in parentheses).

ought to ICE-PHI US ICE-SIN ICE-GBSpoken 7 (4) 22 (6) 33 (20) 105 (63)Written 11 (6) 201 (47) 25 (10) 38 (15)

Frequencies for Figure 30. Frequency of root and epistemic ought to in ICE-PHI, US,ICE-SIN and ICE-GB conversation and fiction (figures normalised to tokens per onemillion words, raw figures in parentheses).

ICE-PHI US ICE-SIN ICE-GBought to Root Root Root Root




Frequencies for Figure 31. Frequency of had better in ICE-PHI, US, ICE-SIN andICE-GB (figures normalised to tokens per one million words, raw figures inparentheses).

had better ICE-PHI US ICE-SIN ICE-GBSpoken 38 (21) 37 (10) 48 (29) 32 (19)Written 13 (5) 33 (33) 25 (10) 20 (8)

102

Frequencies for Figure 32. Frequency of need in ICE-PHI, US, ICE-SIN and ICE-GB(figures normalised to tokens per one million words, raw figures in parentheses).

need ICE-PHI US ICE-SIN ICE-GBSpoken 16 (9) 0 (0) 40 (24) 30 (18)Written 30 (12) 37 (37) 50 (20) 50 (20)

Frequencies for Figure 33. Percentages of positive need and negative need not andneedn’t in ICE-PHI, US, ICE-SIN and ICE-GB.

ICE-PHI need % need not / needn't % Total %Spoken 0 (0) 100 (8) 100 (8)Written 0 (0) 100 (12) 100 (12)

US need % need not / needn't % Total %Spoken 0 (0) 0 (0) 0 (0)Written 24.3 (9) 75.7 (28) 100 (37)

ICE-SIN need % need not / needn't % Total %Spoken 16.7 (4) 83.3 (20) 100 (24)Written 0 (0) 100 (20) 0 (0)ICE-GB need % need not / needn't % Total %Spoken 22.2 (4) 77.8 (14) 100 (18)Written 40 (8) 60 (12) 100 (20)

Frequencies for Figure 34. The frequency of all modal and quasi-modal verbs ofobligation and necessity in ICE-PHI, US, ICE-SIN and ICE-GB (figures normalisedto tokens per one million words, raw figures in parentheses).

ICE-PHI US ICE-SIN ICE-GBSpoken 4169 (2318) 3229 (862) 5262 (3157) 3670 (2202)Written 3175 (1270) 1742 (1742) 3895 (1558) 3545 (1418)

103

Frequencies for Figure 35. The modals and quasi-modals of obligation and necessityin ICE-PHI, US, ICE-SIN and ICE-GB (figures normalised to tokens per one millionwords, raw figures in parentheses). The number of tokens in SBC (the first number inparentheses) has been normalised per 600 000 words and the numbers in Frown per400 000 words. These figures have been added.

ICE-PHI ICE-SIN US ICE-GBhave to 1332 (1273) 1448 (1448) 1293 (472, 582) 1150 (1150)should 1224 (1170) 1447 (1447) 668 (353, 315) 1085 (1085)

must 671 (641) 1023 (1023) 379 (112, 267) 670 (670)need to 189 (181) 364 (364) 417 (112, 63) 256 (256)

supposedto 251 (240) 146 (146) 121 (99, 22) 99 (99)

have got to 26 (25) 174 (174) 38 (22, 16) 217 (217)ought to 11 (10) 30 (30) 32 (13, 19) 78 (78)

had better 27 (26) 39 (39) 35 (22, 13) 27 (27)need 22 (21) 44 (44) 6 (0, 15) 38 (38)

Frequencies for Figure 36. The frequency of should in different text types (figuresnormalised to tokens per one million words, raw figures in parentheses).

ICE-PHI US ICE-SIN ICE-GBPrivate dialogue 1181 (215) 588 (157) 1350 (270) 790 (158)Public dialogue 1650 (231) - 1981 (317) 1206 (193)

Unscripted monologue 992 (133) - 1186 (166) 793 (111)Scripted monologue 1150 (115) - 1360 (136) 1050 (105)

Non-professionalwriting 1300 (52) - 1225 (49) 375 (15)

Correspondence 1067 (64) - 1867 (112) 1683 (101)Academic writing 600 (48) 975 (216) 1063 (85) 675 (54)

Non-academic writing 1225 (98) 711 (199) 713 (57) 975 (78)Reportage 1125 (45) 563 (68) 1450 (58) 1200 (48)

Instructional writing 2475 (99) 760 (55) 2950 (118) 3250 (130)Presuasive writing 2250 (45) 2081 (113) 2400 (48) 2250 (51)Creative writing 675 (27) 545 (137) 775 (31) 1025 (41)

104

Frequencies for Figure 37. The Frequency of supposed to in different text types(figures normalised to tokens per one million words, raw figures in parentheses).

ICE-PHI US ICE-SIN ICE-GBPrivate dialogue 555 (101) 165 (44) 450 (90) 285 (57)Public dialogue 521 (73) - 181 (29) 94 (15)

Unscripted monologue 209 (28) - 36 (5) 50 (7)Scripted monologue 30 (3) - 10 (1) 30 (3)

Non-professionalwriting 0 (0) - 50 (2) 125 (5)

Correspondence 167 (10) - 133 (8) 0Academic writing 0 14 (3) 13 (1) 13 (1)

Non-academic writing 75 (6) 64 (18) 50 (4) 13 (1)Reportage 200 (8) 58 (7) 100 (4) 0 (0)

Instructional writing 125 (5) 41 (3) 0 (0) 75 (3)Presuasive writing 150 (3) 110 (6) 0 (0) 150 (3)Creative writing 75 (3) 72 (18) 50 (2) 100 (4)

Frequencies for Figure 38. Epistemic meaning in ICE-PHI, US, ICE-SIN and ICE-GBconversation (figures normalised to tokens per one million words, raw figures inparentheses).

ICE-PHIEpistemic

USEpistemic

ICE-SINEpistemic

ICE-GBEpistemic

Conversation 198 (36) 302 (21) 415 (83) 420 (84)

Frequencies for Figure 39. Epistemic meaning in ICE-PHI, US, ICE-SIN and ICE-GBfiction (figures normalised to tokens per one million words, raw figures inparentheses).

ICE-PHIEpistemic

USEpistemic

ICE-SINEpistemic

ICE-GBEpistemic

Fiction 525 (21) 364 (85) 450 (18) 900 (36)

Documents

The Modal and Quasi-Modal Verbs of Obligation and